Compare commits

...

338 Commits

Author SHA1 Message Date
zyy17
6e87ac0a0e ci: refine release-cn-artifacts action (#2600)
* ci: add copy-image.sh and upload-artifacts-to-s3.sh

* ci: remove unused options in dev build

* ci: use 'upload-artifacts-to-s3.sh' and 'copy-image.sh' in release-cn-artifacts action

* refactor: refine copy-image.sh
2023-10-13 17:04:06 +08:00
shuiyisong
d89cfd0d4d fix: auth in standalone mode (#2591)
chore: user_provider in standalone mode
2023-10-13 08:37:58 +00:00
Yingwen
8a0054aa89 fix: make nyc-taxi bench work again (#2599)
* fix: invalid requests created by nyc-taxi

* feat: add timestamp to table name

* style: fix clippy

* chore: re-export deps for client

* fix: wait result

* chore: no need to define a prefix constant
2023-10-13 08:16:26 +00:00
Yun Chen
f859932745 fix: convert to ReadableSize & Durations (#2594)
* fix: convert to ReadableSize & Durations

* fix: change more grpc sender/recv message size to ReadableSize

fix: format

fix: cargo fmt

fix: change cmd test to use durations

fix: revert metaclient change

fix: convert default fields in meta client options

fix: human serde meta client durations

* fix: remove milisecond postfix in heartbeat option

* fix: humantime serde on heartbeat

* fix: update config example

* fix: update integration test config

* fix: address pr comments

* fix: fix pr comment on default annotation
2023-10-13 03:28:29 +00:00
Ruihang Xia
9a8fc08e6a docs(benchmark): update 0.4.0 tsbs result (#2597)
* docs(benchmark): update 0.4.0 tsbs result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix typo

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-10-13 03:08:14 +00:00
Ruihang Xia
825e4beead build(ci): pin linux runner to ubuntu-20.04 (#2586)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-10-12 18:08:05 +08:00
zyy17
0a23b40321 ci: downgrade builder version: ubuntu 22.04 -> ubuntu 20.04 for compatible with older version glibc(>=2.31) (#2592) 2023-10-12 16:46:25 +08:00
Ruihang Xia
cf6ef0a30d chore(cli): deregister cli attach command (#2589)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-10-12 08:11:17 +00:00
dennis zhuang
65a659d136 fix: ensure data_home directory created (#2588)
fix: ensure data_home directory created before creating metadata store, #2587
2023-10-12 07:32:55 +00:00
Ruihang Xia
62bcb45787 feat!: change config name from kv_store to metadata_store (#2585)
featchange config name from kv_store to metadata_store

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-10-12 06:55:09 +00:00
zyy17
94f3542a4f ci: fix skopeo running errors (#2581)
ci: fix skopeo auth error
2023-10-12 06:13:56 +00:00
LFC
fc3bc5327d ci: release Windows artifacts (#2574)
* ci: release Windows artifacts

* ci: release Windows artifacts
2023-10-12 14:10:59 +08:00
Ning Sun
9e33ddceea ci: run windows tests every night instead of every commit (#2577)
ci: move windows ci to nightly-ci
2023-10-12 02:53:42 +00:00
zyy17
c9bdf4ff9f ci: refine the process of releasing dev-builder images (#2580)
* fix: fix error of releasing android builder image

* fix: run skopeo error

* ci: add 'release-dev-builder-images-cn' job

* ci: add 'disable_building_images'

* fix: add vars

* ci: use skopeo container

* ci: update opts defaule values
2023-10-12 02:41:54 +00:00
dennis zhuang
0a9972aa9a fix: cache capacity unit in sample config (#2575) 2023-10-11 11:02:39 +00:00
zyy17
76d5b710c8 ci: add more options for releasing dev-builder images (#2573) 2023-10-11 16:24:50 +08:00
zyy17
fe02366ce6 fix: remove unused options and add 'build-android-artifacts' (#2572) 2023-10-11 15:32:58 +08:00
zyy17
d7aeb369a6 refactor: add new action 'release-cn-artifacts' (#2554)
* refactor: add new action 'release-cn-artifacts'

* refactor: refine naming: 'release-artifacts' -> 'publish-github-release'

Signed-off-by: zyy17 <zyylsxm@gmail.com>

---------

Signed-off-by: zyy17 <zyylsxm@gmail.com>
2023-10-11 03:42:04 +00:00
zyy17
9284bb7a2b ci: seperate the job of building dev-builder images (#2569) 2023-10-11 11:09:53 +08:00
liyang
e23dd5a44f fix: fix to readme document link (#2566) 2023-10-11 02:45:43 +00:00
zyy17
c60b59adc8 chore: add the steps of building android binary (#2567) 2023-10-11 02:31:11 +00:00
Lei, HUANG
c9c2b3c91f fix: revert memtable pk rb cache to rwlock (#2565)
* fix: revert memtable pk rb cache to rwlock

* feat: refine
2023-10-10 20:51:05 +08:00
Yingwen
7f75190fce chore: update Cargo.lock (#2564) 2023-10-10 16:28:50 +08:00
Yingwen
0a394c73a2 chore: bump version to 0.4.0 (#2563) 2023-10-10 16:16:15 +08:00
JeremyHi
ae95f23e05 feat: add metrics for region server (#2552)
* feat: add metircs for region server

* fix: add comment and remove unused code
2023-10-10 07:40:16 +00:00
Lei, HUANG
6b39f5923d feat: add compaction metrics (#2560)
* feat: add compaction metrics

* feat: add compaction request total count

* fix: CR comments
2023-10-10 07:38:39 +00:00
JeremyHi
ed725d030f fix: support multi addrs while using etcd (#2562)
fix: support multi addrs while useing etcd
2023-10-10 07:30:48 +00:00
Wei
4fe7e162af fix: human_time mismatch (#2558)
* fix: human_time mismatch.

* fix: add comment
2023-10-10 07:22:12 +00:00
Yingwen
8a5ef826b9 fix(mito): Do not write to memtables if writing wal is failed (#2561)
* feat: add writes total metrics

* fix: don't write memtable if write ctx is failed

* feat: write rows metrics
2023-10-10 06:55:57 +00:00
Ruihang Xia
07be50403e feat: add basic metrics to query (#2559)
* add metrics to merge scan

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* count series in promql

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* tweak label name

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* tweak label name

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* document metric label

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-10-10 06:55:25 +00:00
Lei, HUANG
8bdef9a348 feat: memtable filter push down (#2539)
* feat: memtable support filter pushdown to prune primary keys

* fix: switch to next time series when pk not selected

* fix: allow predicate evaluation failure

* fix: some clippy warnings

* fix: panic when no primary key in schema

* feat: cache decoded record batch for primary key

* refactor: use arcswap instead of rwlock

* fix: format toml
2023-10-10 04:03:10 +00:00
Yingwen
d4577e7372 feat(mito): add metrics to mito engine (#2556)
* feat: allow discarding a timer

* feat: flush metrics

* feat: flush bytes and region count metrics

* refactor: add as_str to get static string

* feat: add handle request elapsed metrics

* feat: add some write related metrics

* style: fix clippy
2023-10-10 03:53:17 +00:00
dennis zhuang
88f26673f0 fix: adds back http_timeout for frontend subcommand (#2555) 2023-10-10 03:05:16 +00:00
Baasit
19f300fc5a feat: renaming kv directory to metadata (#2549)
* fix: renamed kv directory to metadata directory

* fix: changed function name

* fix: changed function name
2023-10-09 11:43:17 +00:00
Weny Xu
cc83764331 fix: check table exists before allocating table id (#2546)
* fix: check table exists before allocating table_id

* chore: apply suggestions from CR
2023-10-09 11:40:10 +00:00
Yingwen
81aa7a4caf chore(mito): change default batch size/row group size (#2550) 2023-10-09 11:10:12 +00:00
Yingwen
d68dd1f3eb fix: schema validation is skipped once we need to fill a column (#2548)
* test: test different order

* test: add tests for missing and invalid columns

* fix: do not skip schema validation while missing columns

* chore: use field_columns()

* test: add tests for different column order
2023-10-09 09:20:51 +00:00
Lei, HUANG
9b3470b049 feat: android image builder dockerfile (#2541)
* feat: android image builder dockerfile

* feat: add building android dev-builder to ci config file

* fix: add build arg

* feat: use makefile to build image and add strip command
2023-10-09 09:10:14 +00:00
Weny Xu
8cc862ff8a refactor: refactor cache invalidator (#2540) 2023-10-09 08:19:18 +00:00
Weny Xu
81ccb58fb4 refactor!: compare with origin bytes during the transactions (#2538)
* refactor: compare with origin bytes during the transaction

* refactor: use serialize_str instead

* Update src/common/meta/src/key.rs

Co-authored-by: JeremyHi <jiachun_feng@proton.me>

* chore: apply suggestions from CR

---------

Co-authored-by: JeremyHi <jiachun_feng@proton.me>
2023-10-09 08:17:19 +00:00
Weny Xu
ce3c10a86e refactor: de/encode protobuf-encoded byte array with base64 (#2545) 2023-10-09 05:31:44 +00:00
shuiyisong
007f7ba03c refactor: extract plugins crate (#2487)
* chore: move frontend plugins fn

* chore: move datanode plugins to fn

* chore: add opt plugins

* chore: add plugins to meta-srv

* chore: setup meta plugins, wait for router extension

* chore: try use configurator for grpc too

* chore: minor fix fmt

* chore: minor fix fmt

* chore: add start meta_srv for hook

* chore: merge develop

* chore: minor fix

* chore: replace Arc<Plugins> with PluginsRef

* chore: fix header

* chore: remove empty file

* chore: modify comments

* chore: remove PluginsRef type alias

* chore: remove `OptPlugins`
2023-10-09 04:54:27 +00:00
Weny Xu
dfe68a7e0b refactor: check push result out of loop (#2511)
* refactor: check push result out of loop

* chore: apply suggestions from CR
2023-10-09 02:49:48 +00:00
Ruihang Xia
d5e4fcaaff feat: dist plan optimize part 2 (#2543)
* allow udf and scalar fn

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* put CountWildcardRule before dist planner

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* bump datafusion to fix first_value/last_value

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update sqlness result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* use retain instead

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-10-09 02:18:36 +00:00
Yingwen
17b385a985 fix: compiler errors under pprof and mem-prof features (#2537)
* fix: compiler errors under pprof feature

* fix: compiler errors under mem-prof feature
2023-10-08 08:28:45 +00:00
shuiyisong
067917845f fix: carry dbname from frontend to datanode (#2520)
* chore: add dbname in region request header for tracking purpose

* chore: fix handle read

* chore: add write meter

* chore: add meter-core to dep

* chore: add converter between RegionRequestHeader and QueryContext & update proto version
2023-10-08 06:30:23 +00:00
Weny Xu
a680133acc feat: enable no delay for mysql, opentsdb, http (#2530)
* refactor: enable no delay for mysql, opentsdb, http

* Apply suggestions from code review

Co-authored-by: Yingwen <realevenyag@gmail.com>

---------

Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-10-08 06:19:52 +00:00
Yingwen
0593c3bde3 fix(mito): pruning for mito2 (#2525)
* fix: pruning for mito2

* chore: refactor projection parameters; add some tests; customize row group size for each flush task.

* chore: pass whole RegionFlushRequest

---------

Co-authored-by: Lei, HUANG <mrsatangel@gmail.com>
2023-10-08 03:45:15 +00:00
Lei, HUANG
0292445476 fix: timestamp range filter (#2533)
* fix: timestamp range filter

* fix: rebase develop

* fix: some style issues
2023-10-08 03:29:02 +00:00
dennis zhuang
ff15bc41d6 feat: improve object storage cache (#2522)
* feat: refactor object storage cache with moka

* chore: minor fixes

* fix: concurrent issues and invalidate cache after write/delete

* chore: minor changes

* fix: cargo lock

* refactor: rename

* chore: change DEFAULT_OBJECT_STORE_CACHE_SIZE to 256Mib

* fix: typo

* chore: style

* fix: toml format

* chore: toml

* fix: toml format

* Update src/object-store/src/layers/lru_cache/read_cache.rs

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>

* chore: update Cargo.toml

Co-authored-by: Yingwen <realevenyag@gmail.com>

* chore: update src/object-store/Cargo.toml

Co-authored-by: Yingwen <realevenyag@gmail.com>

* chore: refactor and apply suggestions

* fix: typo

* feat: adds back allow list for caching

* chore: cr suggestion

Co-authored-by: Yingwen <realevenyag@gmail.com>

* chore: cr suggestion

Co-authored-by: Yingwen <realevenyag@gmail.com>

* refactor: wrap inner Accessor with Arc

* chore: remove run_pending_task in read and write path

* chore: the arc is unnecessary

---------

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-10-08 03:27:49 +00:00
Yingwen
657542c0b8 feat(mito): Cache repeated vector for tags (#2523)
* feat: add vector_cache to CacheManager

* feat: cache repeated vectors

* feat: skip decoding pk if output doesn't contain tags

* test: add TestRegionMetadataBuilder

* test: test ProjectionMapper

* test: test vector cache

* test: test projection mapper convert

* style: fix clippy

* feat: do not cache vector if it is too large

* docs: update comment
2023-10-07 11:36:00 +00:00
Ning Sun
0ad3fb6040 fix: mysql timezone settings (#2534)
* fix: restore time zone settings for mysql

* test: add integration test for time zone

* test: fix unit test for check
2023-10-07 10:21:32 +00:00
Bamboo1
b44e39f897 feat: the schema of RegionMetadata is not output during debug (#2498)
* feat: the schema of RegionMetadata is not output during debug because column_metadatas contains duplicate information

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* feat: the id_to_index of RegionMetadata is not output during debug

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* feat: add debug trait

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* feat: use default debug in ConcreteDataType

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: add std::fmt

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* test: add debug trait test

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: typo

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: resolve conversation

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: test bug

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

---------

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>
2023-10-07 08:01:54 +00:00
Weny Xu
f50f2a84a9 fix: open region missing options (#2473)
* fix: open region missing options

* refactor: remove redundant clone

* chore: apply suggestions from CR

* chore: apply suggestions

* chore: apply suggestions

* test: add test for initialize_region_server

* feat: introduce RegionInfo
2023-10-07 07:17:16 +00:00
Yingwen
fe783c7c1f perf(mito): Use a heap to merge batches for the same key (#2521)
* feat: merge by heap

* fix: fix heap order

* feat: avoid pop/push next and refactor some functions

* feat: replace merge_batches and fixe tests

* test: add test that a key is deleted

* fix: skip empty batch

* style: clippy

* chore: fix typos
2023-10-07 02:56:08 +00:00
Weny Xu
00fe7d104e feat: enable tcp no_delay by default for internal services (#2527) 2023-10-07 02:35:28 +00:00
Zhenchi
201acd152d fix: missing file engine with default options (#2519)
Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-28 10:25:12 +00:00
Niwaka
04dbd835a1 feat: support greatest function (#2490)
* feat: support greatest function

* feat: make greatest take date_type as input

* fix: move sqlness test into common/function/time.sql

* fix: avoid using unwarp

* fix: use downcast

* refactor: simplify arrow cast
2023-09-28 10:25:09 +00:00
Wenjie0329
e3d333258b docs: add event banner (#2518) 2023-09-28 08:08:43 +00:00
Ruihang Xia
10ecc30817 feat: pushdown aggr, limit and sort plan (#2495)
* check partition for aggr plan

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* handle empty partition rule

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove CheckPartition option

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update some valid sqlness result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* opt-out promql plan and update sqlness

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix limit

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix insert select subquery

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update unit test result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix typo

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* Update src/query/src/dist_plan/analyzer.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-09-28 06:35:45 +00:00
JeremyHi
52ac093110 fix: drop table 0 rows affected (#2515) 2023-09-28 06:21:18 +00:00
Zhenchi
1f1d72bdb8 feat: defensively specify limit parameter for file stream (#2517)
Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-28 06:14:27 +00:00
Zhenchi
7edafc3407 feat: push down filters to region engine (#2513)
feat: pushdown filters to region engine

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-27 13:50:44 +00:00
LFC
ccd6de8d6b fix: allow .(dot) literal in table name (#2483)
* fix: allow `.`(dot) literal in table name

* fix: resolve PR comments
2023-09-27 11:50:07 +00:00
shuiyisong
ee8d472aae chore: tune return msg (#2506)
* chore: test return msg

* fix: test_child_error

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* chore: fix test

* chore: minor fix grpc return value

* chore: format return msg

* chore: use root error as return value

* chore: fix empty err display

* chore: iter through external error

* chore: remove err msg

* chore: remove unused field

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-27 10:40:25 +00:00
Weny Xu
9282e59a3b fix: re-create heartbeat stream ASAP (#2499)
* chore: set default connect_timeout_millis to 1000

* fix: re-create heartbeat stream ASAP

* chore: apply suggestions
2023-09-27 04:00:16 +00:00
Ruihang Xia
fbe2f2df46 refactor: simplify warn! and error! macros (#2503)
* refactor: simplify the error! and warn! macros

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* support display format

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* err.msg to err

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-27 03:07:03 +00:00
Yingwen
db6ceda5f0 fix(mito): fix region drop task runs multiple times but never clean the dir (#2504)
fix: fix region drop task runs multiple times but never clean the directory
2023-09-27 02:58:17 +00:00
Ruihang Xia
e352fb4495 fix: check for table scan before expanding (#2491)
* fix: check for table scan before expanding

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* change assert_ok to unwrap

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix clippy warning

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update sqlness result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* don't skip dml

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* uncomment ignored tests

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-26 12:12:08 +00:00
Yingwen
a6116bb866 feat(mito): Add cache manager (#2488)
* feat: add cache manager

* feat: add cache to reader builder

* feat: add AsyncFileReaderCache

* feat: Impl AsyncFileReaderCache

* chore: move moka dep to workspace

* feat: add moka cache to the manager

* feat: implement parquet meta cache

* test: test cache manager

* feat: consider vec size

* style: fix clippy

* test: fix config api test

* feat: divide cache

* test: test disabling meta cache

* test: fix config api test

* feat: remove meta cache if file is purged
2023-09-26 11:46:19 +00:00
Ruihang Xia
515ce825bd feat: stack trace style debug print for error (#2489)
* impl macro stack_trace_debug

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* manually mark external error

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* ignore warnings

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix clippy warnings

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* use debug print

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* simplify the error and warn macro

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix ut

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add docs

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* replace snafu backtrace with location

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-26 11:23:21 +00:00
Vanish
7fc9604735 feat: distribute truncate table in region server (#2414)
* feat: distribute truncate table

* chore: add metrics for truncate table

* test: add sqlness test

* chore: cr

* test: add multi truncate

* chore: add trace id to the header
2023-09-26 11:14:14 +00:00
Zhenchi
a4282415f7 fix: convert datetime to chrono datetime (#2497)
* fix: convert datetime to chrono datetime

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: typo

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix the bad fix

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-26 09:04:12 +00:00
Zhenchi
0bf26642a4 feat: re-support query engine execute dml (#2484)
* feat: re-support query engine execute dml

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: remove region_number in InsertRequest

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: add doc comments

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-26 08:37:04 +00:00
Weny Xu
230a3026ad fix: dn doesn't have chance to send a heartbeat to the new leader (#2471)
* refactor: set meta leader lease secs to 3s

* fix: correct default heartbeat interval

* refactor: ask meta leader in parallel

* feat: configure heartbeat client timeout to 500ms

* fix: trigger to send heartbeat immediately after fail

* fix: fix clippy
2023-09-26 05:05:38 +00:00
Wei
54e506a494 refactor: datetime time unit (#2469)
* refactor: datetime time unit

* Update src/common/time/src/datetime.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

* chore: cr.

---------

Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-09-25 10:00:56 +00:00
Yingwen
7ecfaa240f refactor(mito): remove #[allow(dead_code)] (#2479) 2023-09-25 09:20:00 +00:00
LFC
c0f080df26 fix: print root cause error message to user facing interface (#2486) 2023-09-25 08:44:49 +00:00
Niwaka
f9351e4fb5 chore: add integration test for issue2437 (#2481) 2023-09-25 06:23:16 +00:00
zyy17
00272d53cc chore: fix typo (#2477) 2023-09-24 06:47:14 +00:00
JeremyHi
7310ec0bb3 chore: refactor options (#2476) 2023-09-24 02:12:33 +00:00
Yingwen
73842f10e7 fix(mito): normalize region dir in RegionOpener (#2475)
fix: normalize region dir in RegionOpener
2023-09-23 10:06:00 +00:00
Yingwen
32d1d68441 fix(mito): reset is_sorted to true after the merger finishing one series (#2474)
fix: reset is_sorted flag to true after the merger finishing one series
2023-09-23 10:05:34 +00:00
Ning Sun
ffa729cdf5 feat: implement storage for OTLP histogram (#2282)
* feat: implement new histogram data model

* feat:  use prometheus table format for histogram

* refactor: remove duplicated code

* fix: histogram tag column

* fix: use accumulated count in buckets

* refactor: using row based protocol for otlp WIP

* refactor: use row based writer for otlp.

Also updated row writer for owned keys

* refactor: use row writers for otlp

* test: add integration tests for histogram

* refactor: change le column name
2023-09-23 07:59:14 +00:00
JeremyHi
9d0de25bff chore: typo (#2470) 2023-09-22 09:47:34 +00:00
Wei
aef9e7bfc3 refactor: not allowed int64 type as time index (#2460)
* refactor: remove is_timestamp_compatible.

* chore: fmt

* refactor: remove int64 to timestamp match

* chore

* chore: apply suggestions from code review

Co-authored-by: dennis zhuang <killme2008@gmail.com>

* chore: fmt

---------

Co-authored-by: dennis zhuang <killme2008@gmail.com>
2023-09-22 06:28:02 +00:00
Yingwen
c6e95ffe63 fix(mito): compaction scheduler schedules more tasks than expected (#2466)
* test: test on_compaction_finished

* fix: avoid submit same region to compact

* feat: persist and recover compaction time window

* test: fix test

* test: sort like result
2023-09-22 06:13:12 +00:00
Yingwen
c9f8b9c7c3 feat: update proto and remove create_if_not_exists (#2467) 2023-09-22 03:24:49 +00:00
Baasit
688e64632d feat: support for show full tables (#2410)
* feat: added show tables command

* fix(tests): fixed parser and statement unit tests

* chore: implemeted display trait for table type

* fix: handled no tabletype and error for usopprted command in show databse

* chore: removed full as a show kind, instead as a show option

* chore(tests): fixed failing test and added more tests for show full

* chore: refactored table types to use filters

* fix: changed table_type to tables
2023-09-22 02:34:57 +00:00
JeremyHi
8e5eaf5472 chore: remove unused region_stats method form table (#2458)
chore: remove unused region_status method form table
2023-09-22 02:27:29 +00:00
LinFeng
621c6f371b feat: limit grpc message size (#2459)
* feat: add two grpc config options

Those options are for:
* Limit receiving(decoding) message size
* Limit sending(enoding) message size

* test: add integration tests for message size limit
2023-09-22 02:07:46 +00:00
Ruihang Xia
4c7ad44605 refactor: remove SqlStatementExecutor (#2464)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-22 01:57:48 +00:00
Weny Xu
6306aeabf0 chore: bump opendal to 0.40 (#2465) 2023-09-21 14:25:23 +00:00
JeremyHi
40781ec754 fix: test on windows (#2462)
* fix: test on windows

* fix: fix windows root

* fix: use relative path instead of root

* fix: remove incorrect replace

* fix: fix tests

---------

Co-authored-by: WenyXu <wenymedia@gmail.com>
2023-09-21 10:57:56 +00:00
zyy17
c7b490e1a0 ci: expand upload retry timeout (#2461) 2023-09-21 10:02:15 +00:00
Ruihang Xia
e3f53a8060 fix: add slash after generated region_dir (#2463)
* fix: add slash after generated region_dir

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update ut

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-21 07:46:05 +00:00
Weny Xu
580d11b1e1 refactor: unify table metadata cache invalidator (#2449)
* refactor: unify table metadata cache invalidator

* chore: apply from suggestions
2023-09-21 03:45:49 +00:00
shuiyisong
20f4f7971a refactor: remove source and location in snafu display (#2428)
* refactor: remove source pt 1

* refactor: remove source pt 2

* refactor: remove source pt 3

* refactor: remove location pt 1

* refactor: remove location pt 2

* chore: remove rustc files

* chore: fix error case

* chore: fix test

* chore: fix test

* chore: fix cr issue

Co-authored-by: fys <40801205+fengys1996@users.noreply.github.com>

---------

Co-authored-by: fys <40801205+fengys1996@users.noreply.github.com>
2023-09-21 02:55:24 +00:00
dennis zhuang
9863e501f1 test: revert ignored tests (#2455) 2023-09-21 02:33:18 +00:00
zyy17
df0877111e ci: make upload-to-s3 configurable(for now, it's false) (#2456) 2023-09-20 14:12:54 +00:00
dennis zhuang
23cc7d82e5 feat: supports binary data type (#2454) 2023-09-20 12:53:19 +00:00
Ruihang Xia
34d6288945 feat: bring back sqlness and integration tests (#2448)
* enable integration test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update sqlness result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* disable sqlness region failover

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* enable sqlness in CI

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* sort unstable result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* set require_lease_before_startup to true

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix: fix inconsistent cache

* replace windows path chars

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* ignore some integration cases in windows

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* Revert "ignore some integration cases in windows"

This reverts commit 122478b7c1.

* disable windows for now

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix: fix close region bug in RegionHeartbeatResponseHandler

* disable failover tests

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: WenyXu <wenymedia@gmail.com>
2023-09-20 09:17:30 +00:00
JeremyHi
567fbad647 chore: type alias typo (#2452)
chore: typo
2023-09-20 07:53:35 +00:00
Ruihang Xia
a5c499572c feat: open region in parallel (#2451)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-20 07:40:17 +00:00
JeremyHi
ca50ba5dc4 fix: remark region as inactive on leader changed (#2446)
* fix: remark reigon as inactive on leader changed

* chore: by comment
2023-09-20 06:37:27 +00:00
Yingwen
17e560c909 feat(mito): Allow to retry create request and alter request (#2447)
* feat: RegionMetadataBuilder allow adding/dropping columns multiple times

* test: test add if not exists/drop if exists

* feat: change validator and add need_alter

* test: fix tests and test need_alter

* test: test alter retry

* feat: open before create

* style: fix clippy
2023-09-20 06:36:46 +00:00
Weny Xu
339e12c64a fix: fix alter table verification (#2437)
* fix: fix verify alter

* refactor: move AlterTable UpdateMetadata to last step

* refactor: send region request in parallel

* Update src/table/src/metadata.rs

Co-authored-by: LFC <990479+MichaelScofield@users.noreply.github.com>

* Update src/table/src/metadata.rs

Co-authored-by: JeremyHi <jiachun_feng@proton.me>

---------

Co-authored-by: LFC <990479+MichaelScofield@users.noreply.github.com>
Co-authored-by: JeremyHi <jiachun_feng@proton.me>
2023-09-19 13:40:48 +00:00
Ruihang Xia
0f79ccab31 refactor: remove the old mito engine (#2443)
Co-authored-by: Even Yag <realevenyag@gmail.com>
2023-09-19 09:30:13 +00:00
Yingwen
7b606ed289 feat(mito): make use of options in RegionCreate/OpenRequest (#2436)
* refactor: move RegionOptions to options mod

* refactor: define compaction strategy in region/options.rs

* feat: use duration for time window

* refactor: rename CompactionStrategy to CompactionOptions

* feat: use serde to parse options

* feat: parse options

* feat: set options on creation/opening

* test: test create/open with options

* chore: remove todo

* feat: get compaction ttl and options from RegionOptions

* style: fix clippy

* chore: Remove unused engine_options

* style: fix clippy

* chore: remove todo
2023-09-19 09:06:09 +00:00
Weny Xu
1fb2d95c5f fix: fix open region missing path (#2441)
* fix: fix open region missing path

* fix: correct log

* chore: apply suggestions from CR

* fix: fix tests
2023-09-19 08:50:59 +00:00
Wei
8ee62a7d90 fix: parse i64 to different kinds of timestamp (#2440)
* feat: support i64 to multi-timestamp.

* chore: fmt
2023-09-19 08:26:25 +00:00
Ruihang Xia
802229de87 fix: type cast bugs found by sqlness (#2438)
* update valid results

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* accomplish datatype

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* cast null

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix unit tests

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-19 08:20:41 +00:00
Zhenchi
deac284973 refactor: RegionRequestHandler -> RegionQueryHandler (#2439)
* refactor: RegionRequestHandler -> RegionQueryHandler

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: rename FrontendRegionQueryHandler

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: privte RegionInvoker

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-19 08:19:58 +00:00
Wei
5805e8d4b6 feat: type conversion between Values (#2394)
* feat: add cast() in datatype trait.

* feat: add cast for primitive type

* feat: add unit test cases

* test: add datetime/time cases.

* refactor: time_type cast function.

* chore: typos.

* refactor code.

* feat: add can_cast_type func.

* chore: rename cast to try_cast

* feat: impl cast_with_opt

* chore: pub use cast_with_opt

* chore: add timezone for test

* Update src/common/time/src/date.rs

Co-authored-by: dennis zhuang <killme2008@gmail.com>

* chore: duration type

---------

Co-authored-by: dennis zhuang <killme2008@gmail.com>
2023-09-18 14:25:38 +00:00
dennis zhuang
342cc0a4c4 fix: compile error after updating protos (#2435) 2023-09-18 12:12:39 +00:00
Weny Xu
df6c79a378 fix: check version before alter region (#2433)
* fix: check version before alter region

* chore: apply suggestions from CR

* Update src/mito2/src/worker/handle_alter.rs

Co-authored-by: dennis zhuang <killme2008@gmail.com>

---------

Co-authored-by: dennis zhuang <killme2008@gmail.com>
2023-09-18 11:49:26 +00:00
dennis zhuang
5566f34bd1 feat: make scripts table work again (#2420)
* feat: make scripts table work again

* chore: typo

* fix: license header

* Update src/table/src/metadata.rs

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>

* chore: cr comments

---------

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-18 11:43:21 +00:00
Wei
14e6998d41 feat: impl duration datatype and vectors (#2180)
* feat: impl datatype, vector traits for duration.

* feat: duration and grpc.

* test: add unit test cases.

* chore: style and test case.

* fix: update greptime-proto version and helper.rs

* chore: fix type name.

* Update src/datatypes/src/data_type.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

* chore: cr.

* chore: fix greptime-proto

---------

Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-09-18 11:28:06 +00:00
Weny Xu
43476e1ff9 refactor: rename coordination to require_lease_before_startup (#2431) 2023-09-18 11:07:42 +00:00
Weny Xu
c42cce57ca fix: fix incorrect matches (#2430)
* fix: fix incorrect matches

* fix: fix incorrect status code
2023-09-18 10:53:32 +00:00
Zhenchi
d6d46378a1 test: fix some integration tests (#2432)
* test: fix some integration tests

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* test: add timezone setting

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-18 10:52:14 +00:00
Ruihang Xia
fbbf3978d9 fix: render comment in SHOW CREATE TABLE (#2427)
* feat: add comment field to ColumnDef

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix sqlness case

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-18 10:51:10 +00:00
dennis zhuang
b0c56a3e23 feat: type alias (#2331)
* fix: remove location from error msg

* feat: adds transformer for sqlparser statements

* feat: supports type alias

* fix: typo

* fix: license header

* test: adds timestamp_types test

* refactor: transform

* fix: rebase develop and fix tests

* fix: compile error

* chore: delete src/datanode/src/sql/create_external.rs
2023-09-18 09:43:02 +00:00
Zhenchi
73af1368bd test: more integration test cases for external table (#2429)
Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-18 09:21:51 +00:00
Weny Xu
2c3ff90dbc feat: start services after first heartbeat response processed (#2424)
* feat: start services after first heartbeat response processed

* refactor: watch changes in RegionAliveKeeper

* feat: add coordination to DatanodeOptions

* chore: apply suggestions from CR

* chore: apply suggestions from CR

* chore: enable coordination in sqlness
2023-09-18 08:49:26 +00:00
Zhenchi
3a39215f11 feat: migrate file engine from table to reigon (#2365)
* feat: migrate file engine from table to reigon

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* Update src/file-engine/src/engine.rs

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>

* feat: specify ts index for file engine

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: handle time index for external table

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: some integration testsg

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: add file schema and table schema compatibility

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: compatible file schema to region schema

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: add error msg

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: simplify close

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: implement set_writable

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: tests-integration compilation

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: address comments

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: zhongzc <zhongzc@zhongzcs-MacBook-Pro.local>
2023-09-18 08:02:43 +00:00
JeremyHi
e7e254cd11 feat: all distributed time together (#2423) 2023-09-17 15:18:52 +00:00
Yingwen
4a82926d72 docs: fix cargo doc errors and warnings (#2421)
* docs: fix cargo doc warnings and errors

* docs: fix warnings

* docs: fix warnings

* chore: rm src/common/function-macro/src/lib.rs
2023-09-17 11:45:15 +00:00
Ruihang Xia
92824d1c66 fix: update several sqlness results (#2422)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-17 11:33:40 +00:00
Yingwen
55ae5e5b66 feat(mito): Implements compaction scheduler (#2413)
* feat: allow multiple waiters in compaction request

* feat: compaction status wip

* feat: track region status in compaction scheduler

* feat: impl compaction scheduler

* feat: call compaction scheduler

* feat: remove status if nothing to compact

* feat: schedule compaction after flush

* feat: set compacting to false after compaction finished

* refactor: flush status only needs region id and version control

* refactor: schedule_compaction don't need region as argument

* test: test flush/scheduler for empty requests

* test: trigger compaction in test

* feat: notify scheduler on truncated

* chore: Apply suggestions from code review

Co-authored-by: JeremyHi <jiachun_feng@proton.me>

---------

Co-authored-by: JeremyHi <jiachun_feng@proton.me>
2023-09-17 09:15:11 +00:00
Ruihang Xia
693e8de83a feat: scope spawned task with trace id (#2419)
* feat: scope spawned task with trace id

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-17 09:05:28 +00:00
JeremyHi
542e863ecc fix: missing datanode id on keep lease (#2415) 2023-09-17 07:57:17 +00:00
Ruihang Xia
49310acea1 refactor: rename common-function-macro subcrate (#2418)
* rename common-function-macro to common-macro

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* put impl into their own file

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-17 07:56:41 +00:00
Weny Xu
5b08e03944 feat: sync regions between RegionServer and RegionAliveKeeper (#2417)
* feat: sync regions between RegionServer and RegionAliveKeeper

* Apply suggestions from code review

Co-authored-by: JeremyHi <jiachun_feng@proton.me>

* refactor: rename event name

* chore: apply suggestions

---------

Co-authored-by: JeremyHi <jiachun_feng@proton.me>
2023-09-17 07:55:44 +00:00
JeremyHi
98a40bae95 feat!: unify naming with options (#2416) 2023-09-17 07:24:57 +00:00
JeremyHi
342a6d071f feat: heartbeat request with header (#2412)
* feat: heartbeat request with header

* chore: frontend send heartbeat with a longer interval
2023-09-16 09:56:41 +00:00
Baasit
0a692aafb0 feat: clap wrapper around sqlness (#2400)
* feat: wrapped sqlness with clap to provide nice interface

* fix: added spaces and changed -f flag to bool
2023-09-16 08:53:08 +00:00
dennis zhuang
627c5b7419 feat: move table operations from frontend to operator crate (#2411)
* feat: move table operations from frontend to operator crate

* chore: blank line

* fix: toml format

* chore: move constants
2023-09-16 07:58:45 +00:00
Ruihang Xia
5e35087b67 fix: generate region path with given prefix (#2409)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-16 03:23:36 +00:00
Ruihang Xia
c149c123c3 feat: reopen corresponding regions on starting datanode (#2399)
* separate config and datanode impl

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* partial implement of fetching region id list

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* reopen all regions on starting region server

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update sqlness & assign default datanode id

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* set writable on lease

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* apply cr suggs.

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* Update src/datanode/src/datanode.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: dennis zhuang <killme2008@gmail.com>
Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-09-15 13:30:20 +00:00
Vanish
0bd6b9bb39 feat: implement truncate region for mito2 (#2335)
* feat: implement truncate region for mito2.

* chore: add license header and fix typos

* Update src/mito2/src/worker/handle_truncate.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

* cr

* chore: consider the flush task being executed before truncating the region.

* test

* feat: check flush and compaction tasks

* chore: remove useless changes

* Update src/mito2/src/manifest/action.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

* Update src/mito2/src/worker/handle_flush.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

* chore: CR, consider sequence number

* test: use EventListener to test the flush task during truncate

* fix: fix listener error

* Update src/mito2/src/engine/truncate_test.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

* chore: cr

* fix: remove set None

* Update src/mito2/src/region/version.rs

Co-authored-by: dennis zhuang <killme2008@gmail.com>

* Update src/mito2/src/worker/handle_flush.rs

Co-authored-by: dennis zhuang <killme2008@gmail.com>

* Update src/mito2/src/worker/handle_truncate.rs

Co-authored-by: dennis zhuang <killme2008@gmail.com>

* doc: add some doc for FlushTruncateListener and RegionTruncate

---------

Co-authored-by: Yingwen <realevenyag@gmail.com>
Co-authored-by: dennis zhuang <killme2008@gmail.com>
2023-09-15 13:20:01 +00:00
Bamboo1
6aec30a1a8 feat: reserve internal column (#2396)
* feat: reserve internal column

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* test: add function test

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: spell typos

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: resolve conversation

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

---------

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>
2023-09-15 11:19:40 +00:00
LFC
a688760563 fix: validate partition columns (#2393)
fix: partition column must belong to primary keys or equals to time index
2023-09-15 10:07:32 +00:00
LFC
4b13c88752 fix: resolve more integration tests (#2406)
* fix: resolve more integration tests

* Update tests-integration/tests/http.rs

Co-authored-by: Weny Xu <wenymedia@gmail.com>

---------

Co-authored-by: Weny Xu <wenymedia@gmail.com>
2023-09-15 09:43:16 +00:00
JeremyHi
9572b1edbb feat: region storage path (#2404)
* feat: region storage path

* Update src/common/meta/src/key/datanode_table.rs

Co-authored-by: Weny Xu <wenymedia@gmail.com>

* chore: by cr

* feat: upgrade proto

---------

Co-authored-by: Weny Xu <wenymedia@gmail.com>
2023-09-15 09:07:54 +00:00
dennis zhuang
43e3c94fd1 refactor: catalog managers (#2405)
* feat: rename catalog::local to catalog::memory

* refactor: catalog managers

* chore: license header
2023-09-15 08:48:14 +00:00
Ruihang Xia
364b99a14c fix: enable ignored promql unit tests (#2403)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-15 07:37:30 +00:00
shuiyisong
a8ae386a57 chore: add #[serde(default)] to new added engine field (#2402)
chore: add serde default to new field
2023-09-15 07:11:57 +00:00
LFC
fe5679e77e refactor: remove table ident (#2368)
* refactor:
1. remove TableIdent, use TableId directly
2. use the latest greptime-proto
3. independently invalidate table id cache and table name cache

* rebase

* fix: resolve PR comments

* fix: resolve PR comments
2023-09-15 05:14:40 +00:00
JeremyHi
8e70b9e982 feat: remove deprecated metadata keys (#2398)
* feat: remove deprecated metadata keys

* feat: this time, weny indeed said [removes it]
2023-09-15 02:11:21 +00:00
JeremyHi
d1adb915bf feat: set readonly first when deregister region (#2391)
* feat: set readonly first when deregister region

* revert distxxx
2023-09-14 12:12:38 +00:00
Yingwen
a84a8ad04f fix: alter table procedure panics while renaming table (#2397)
* fix: procedure panic on renaming table

* test: fix test_insert_and_select invalid arguments

* test: fix test_standalone_insert_and_query using wrong semantic type

* test: fix test_distributed_insert_delete_and_query semantic type
2023-09-14 11:50:00 +00:00
JeremyHi
7bb8a5999c feat!: add engine name to DatanodeTableValue (#2395)
* feat: add engine name to DatanodeTableValue

* fix: by cr
2023-09-14 09:50:35 +00:00
Yingwen
26992d58cd chore: decrease mutable write buffer limit (#2390)
* chore: set mutable limit to half of the global write buffer size

* refactor: put handle_flush_finished after handle_flush_request

* refactor: rename tests.rs to basic_test.rs

* style: fmt code
2023-09-14 08:24:14 +00:00
Ruihang Xia
47bf300869 fix: update sqlness result for order_by (#2389)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-14 07:28:40 +00:00
Yingwen
a7df5a7c9a fix(mito): incorrect field index in ProjectionMapper (#2388)
* chore: update todo comments

* test: add test for projection

* fix: panics when projecting fields

* chore: remove todos
2023-09-14 04:15:15 +00:00
Yingwen
d4ae8a6fed feat(mito): Add writable flag to region (#2349)
* feat: add writable flag to region.

* refactor: rename MitoEngine to MitoEngine::scanner

* feat: add set_writable() to RegionEngine

* feat: check whether region is writable

* feat: make set_writable sync

* test: test set_writable

* docs: update comments

* feat: send result on compaction failure

* refactor: wrap output sender in new type

* feat: on failure

* refactor: use get_region_or/writable_region_or

* refactor: remove send_result

* feat: notify waiters on flush scheduler drop

* test: fix tests

* fix: only alter writable region
2023-09-14 02:45:30 +00:00
Yingwen
da54a0c139 fix: alter table procedure forgets to update next column id (#2385)
* feat: add more info to error messages

* feat: store next column id in procedure

* fix: update next column id for table info

* test: fix add col test

* chore: remove location from invalid request error

* test: update test

* test: fix test
2023-09-14 02:06:57 +00:00
Ruihang Xia
cc7eb3d317 fix: querying temporary table (#2387)
* fix information schema

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove log

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-13 14:30:56 +00:00
Weny Xu
93f3048f4f refactor: migrate OpenDal to 0.39 (#2383)
* chore: bump opendal to 7d552

* refactor: migrate OpenDal to 0.39

* chore: apply suggestions from CR
2023-09-13 12:43:53 +00:00
LFC
d08b05c963 fix: make test-integration able to compile (#2384)
* fix: make test-integration able to compile

* chore: fmt toml

---------

Co-authored-by: WenyXu <wenymedia@gmail.com>
2023-09-13 12:42:46 +00:00
JeremyHi
f76aa278fd feat: atomic metadata (#2366)
* feat: atomic creating metadata

* chore: exist exists

* chore: license header

* chore: weny never say that

* feat: add put_conditionally to kv_backend
2023-09-13 10:51:05 +00:00
JeremyHi
6f4779b474 feat: engine name in heartbeat (#2377) 2023-09-13 09:10:10 +00:00
Ruihang Xia
de723d9c1c fix: update sqlness result in distributed mode (#2381)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-13 09:07:55 +00:00
dennis zhuang
7448e975c2 chore: change error messages (#2379)
* chore: change error messages

* chore: remove locaton in table not found error msg
2023-09-13 08:21:03 +00:00
dennis zhuang
3f97a0d285 fix: gRPC max mesage size limitation (#2375)
* fix: gRPC max mesage size limitation

* chore: don't set max_encoding_message_size
2023-09-13 08:13:49 +00:00
Ruihang Xia
60bdf9685f feat: use the latest command line options for sqlness runner (#2371)
feat: use the latest command line options

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-13 03:38:43 +00:00
Ruihang Xia
9c76d2cf54 feat: convert sql number to values with target type (#2370)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-13 11:14:42 +08:00
Weny Xu
1a7268186b chore: bump raft-engine to 22dfb4 (#2360) 2023-09-12 07:57:15 -05:00
Ruihang Xia
eeecce4623 refactor: remove table procedure (#2359)
remove table procedure

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
Ruihang Xia
1ad5f6e5d5 refactor: system tables in FrontendCatalogManager (#2358)
* rename method names

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove system table, table engine, register/deregister

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add system catalog

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* run nextest

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* some documents

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix: fix clippy

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: WenyXu <wenymedia@gmail.com>
2023-09-12 07:57:15 -05:00
Yingwen
46eca5026e fix(mito): Stores and recovers flushed sequence (#2355)
* test: add test for reopen

* feat: last entry id starts from flushed entry id

* fix: store flushed sequence and recover it from manifest

* test: check sequence in alter test

* test: more tests for alter
2023-09-12 07:57:15 -05:00
Weny Xu
912341e4fa fix: fix start issues under standalone mode (#2352)
* fix: fix standalone starts

* chore: bump raft-engine to 571462e

* refactor: remove MetadataService
2023-09-12 07:57:15 -05:00
JeremyHi
80c5d52015 feat: stop region server (#2356)
* feat: stop region server

* fix: close region first
2023-09-12 07:57:15 -05:00
Zhenchi
4af126eb1b feat: consolidate Insert request related partitioning and distributed processing operations into Inserter (#2346)
* refactor: RegionRequest as param of RegionRequestHandler.handle

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: partition insert & delete reqs for both standalone and distributed mode

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: nit change

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: wrong function nameg

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: do request in inserter & deleter

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: remove RegionRequestHandler.handle

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: rename table_creator

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: nit change

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: address comments

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: nit change

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-12 07:57:15 -05:00
LFC
fe954b78a2 refactor: system tables in new region server (#2344)
refactor: inverse the dependency between system tables and catalog manager
2023-09-12 07:57:15 -05:00
JeremyHi
3cab6de391 feat: filter out empty heartbeat req (#2345)
* feat: filter out empty heartbeat request

* fix: big mistake
2023-09-12 07:57:15 -05:00
Yingwen
606ee43f1d feat(mito): Implement skeleton for alteration (#2343)
* feat: impl handle_alter wip

* refactor: move send_result to worker.rs

* feat: skeleton for handle_alter_request

* feat: write requests should wait for alteration

* feat: define alter request

* chore: no warnings

* fix: remove memtables after flush

* chore: update comments and impl add_write_request_to_pending

* feat: add schema version to RegionMetadata

* feat: impl alter_schema/can_alter_directly

* chore: use send_result

* test: pull next_batch again

* feat: convert pb AlterRequest to RegionAlterRequest

* feat: validate alter request

* feat: validate request and alter metadata

* feat: allow none location

* test: test alter

* fix: recover files and flushed entry id from manifest

* test: test alter

* chore: change comments and variables

* chore: fix compiler errors

* feat: add is_empty() to MemtableVersion

* test: fix metadata alter test

* fix: Compaction picker doesn't notify waiters if it returns None

* chore: address CR comments

* test: add tests for alter request

* refactor: use send_result
2023-09-12 07:57:15 -05:00
Lei, HUANG
3331e3158c feat(mito2): compaction (#2317)
* feat: compaction component

* feat: mito2 compaction

* Avoid building time range predicates when merge SST files since in TWCS we don't enforce strict time window.

* fix: some CR comments

* minor: change CompactionRequest::senders to an option

* chore: handle compaction finish error

* feat: integrate compaction into region worker

* chore: rebase upstream

* fix: Some CR comments

* chore: Apply suggestions from code review

* style: fix clippy

---------

Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-09-12 07:57:15 -05:00
Weny Xu
a4604afde5 refactor: rename NEXT_TABLE_ROUTE_PREFIX to TABLE_ROUTE_PREFIX (#2348)
* refactor: rename NEXT_TABLE_ROUTE_PREFIX to TABLE_ROUTE_PREFIX

* chore: apply suggestions from CR
2023-09-12 07:57:15 -05:00
Weny Xu
f386329e29 refactor: introduce DdlTaskExecutor and refactor statement executor (#2341)
* feat: add kv store option

* refactor: refactor statement executor

* refactor: refactor standalone table creator

* chore: apply suggestions from CR

* chore: apply suggestions from CR

* refactor: move ShowCreateTable and CreateDatabase to StatementExecutor

* fix: fix RegionDistribution

* feat: build standalone

* chore: apply suggestions from CR

* chore: apply suggestions from CR

* chore: apply suggestions from CR
2023-09-12 07:57:15 -05:00
Yingwen
3f6d557b8d feat: Implements a reader to make schema compatible (#2326)
* docs: update comment

* feat: Add compat reader to SeqScan

* feat: add struct to compat pk and fields

* refactor: remove unused fields from ParquetReader

* feat: compat framework

* feat: Implement CompatPrimaryKey and CompatFields

* feat: implement compat reader

* feat: Test compat reader

* test: test compat reader

* feat: add more checks to concat

* style: fix clippy

* test: more tests for compat reader

* test: test reader with projection
2023-09-12 07:57:15 -05:00
Ruihang Xia
6215f124f7 refactor: remove datanode instance (#2342)
* pass nextest

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove deadcode

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* rename region_alive_keepers

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
LFC
1d83c942a9 refactor: script table creation (#2340)
* refactor:
1. remove method `register_system_table` from CatalogManager
2. the creation of ScriptTable (as a system table) is removed from CatalogManager. Instead, the ScriptTable is created when Frontend instance is starting; and is created by calling Frontend instance's grpc handler.

* rebase
2023-09-12 07:57:15 -05:00
Ruihang Xia
f287a5db9f feat: adapt region keep aliver for region server (#2333)
* basic impl

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* refactor, collapse one layer

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove old heartbeat handler impls

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove old region alive keeper

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove remote catalog manager

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* global replace

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* test countdown task

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
Zhenchi
dac6b2e80a feat(frontend): migrate delete to region server (#2329)
* feat(frontend): migrate delete to region server

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: add more check and do trim columns

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: RegionRequestHandler.handle retrun AffectedRows

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-12 07:57:15 -05:00
Yingwen
1e44e86d81 feat(mito): Stall write requests and add more flush tests (#2322)
* feat: impl reject write

* feat: sanitize reject size

* feat: add should_stall to WriteBufferManager

* feat: stall requests

* test: mock WriteBufferManager

* feat: add new_with_manager for test and remove object_store from inner

* feat: add an event listener for tests

* feat: Use listener to test flush

* refactor: add flush_test.rs

* style: fix clippy

* feat: test write stall

* test: test flush empty
2023-09-12 07:57:15 -05:00
JeremyHi
56691ff03b refactor: mailbox timeout (#2330)
refactor: Optimize the timeout mechanism of the mailbox
2023-09-12 07:57:15 -05:00
Weny Xu
e4de63625f refactor: refactor raft engine backend and state store (#2336)
* refactor: remove redundant code

* refactor: refactor RaftEngineBackend Error to common_meta::error::Error

* refactor: refactor state store

* chore: apply suggestions from CR
2023-09-12 07:57:15 -05:00
Ruihang Xia
4b2b59c31b refactor: clean unnecessary disabled lints (#2338)
* clean manifest

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean engine

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean region

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean asscess_layer

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean manifest manager

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean row_converter

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean scheduler

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean worker

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
Weny Xu
2ee2d29085 refactor: move Sequence to common meta (#2337) 2023-09-12 07:57:15 -05:00
Yingwen
c3f6529178 fix: improve error message in validate_proto_value (#2328)
* fix: correct error message in validate_proto_value()

* fix: print location in InvalidRequest error

* style: format
2023-09-12 07:57:15 -05:00
Ruihang Xia
eb7116ab56 feat: read/write works in distributed mode 🎉 (#2327)
* add do_get method to RegionRequestHandler

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* move RegionRequestHandler to client crate

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* use RegionRequestHandler in MergeScan

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* minor fix

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* ignore tests

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix format

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
Zhenchi
5f7d48f107 feat(frontend): reorg insert converters and introduce stmt_to_region (#2324)
* feat(frontend): reorg insert converters and introduce stmt_to_region

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: shorten import path

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: add check for column count

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: clippy

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-12 07:57:15 -05:00
LFC
711e27d9fa feat: distributed alter table in region server (#2311)
* feat: distributed alter table in region server

* rebase
2023-09-12 07:57:15 -05:00
Weny Xu
922e342b63 refactor: refactor ddl manager (#2306)
* refactor: refactor ddl manager

* chore: apply suggestions from CR
2023-09-12 07:57:15 -05:00
Zhenchi
7dde9ce3ce feat(frontend): migrate insert to region server (#2318)
* feat(frontend): migrate insert to region server

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: move converter to Inserter

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: rename convert function

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: address comments

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: address comments

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: add span id

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: compilation

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* retrigger action

* retrigger action

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-12 07:57:15 -05:00
Yingwen
3eccb36047 feat: avoid using vector to get default value (#2323) 2023-09-12 07:57:15 -05:00
Ruihang Xia
f71aa373c1 feat: start datanode with config (#2312)
* remove memory-catalog and procedure

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* derive serde for MitoConfig

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* start datanode with configs

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove dir in WalConfig

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add rename field attr

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add stupid duplicated mito config

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove wrong import

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* wired compile error

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
Ruihang Xia
50fca2400e feat: adapt methods from RegionEngine for MitoEngine (#2315)
* feat: adapt methods from RegionEngine for MitoEngine

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* minor fixes

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
JeremyHi
920763d7dd feat: add metric and manage tool for InactiveRegionKey (#2313)
* feat: add metric and manage tool for InactiveRegionKey

* chore: by review comment
2023-09-12 07:57:15 -05:00
dennis zhuang
a3d5931fca feat: unify all protocol options (#2316)
* feat: unify all protocol options

* feat: adds enable to example configs

* chore: style

Co-authored-by: JeremyHi <jiachun_feng@proton.me>

---------

Co-authored-by: JeremyHi <jiachun_feng@proton.me>
2023-09-12 07:57:15 -05:00
dennis zhuang
b1599ad3a5 fix: can't adding new columns as primary key (#2310) 2023-09-12 07:57:15 -05:00
dennis zhuang
38697e0c4d feat: build http client for cloud object storage (#2314)
* feat: build http client for s3/oss/azblob storages

* chore: style

* fix: test

* fix: cargo toml fmt
2023-09-12 07:57:15 -05:00
Yingwen
50220f8f04 feat: Impl write buffer manager for mito2 (#2309)
* feat: add write buffer manager to builder

* feat: impl WriteBufferManager

* feat: impl MemtableVersion::mutable_usage

* chore: Address CR comments

Co-authored-by: JeremyHi <jiachun_feng@proton.me>

* refactor: rename mutable_limitation to mutable_limit

---------

Co-authored-by: JeremyHi <jiachun_feng@proton.me>
2023-09-12 07:57:15 -05:00
Niwaka
3504d8254e fix: unused table options (#2267)
* fix: unused table options keys

* refactor: simplify validate table options

* chore: Add newlines

---------

Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-09-12 07:57:15 -05:00
dennis zhuang
fad58835bf fix: don't raise an error when manifest directory is not created (#2308)
* fix: don't raise an error when manifest directory is not created

* chore: apply suggestion

Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>

---------

Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>
2023-09-12 07:57:15 -05:00
Lei, HUANG
43fdff3639 feat: remove memtable request (#2307)
* refactor: remove scan request from memtable API

* docs: Update comment

---------

Co-authored-by: Yingwen <realevenyag@gmail.com>
2023-09-12 07:57:15 -05:00
Weny Xu
271f80daad fix: LoadBase Selector cannot follow the region distribution rules (#2259)
* fix: LoadBase Selector cannot follow the region distribution rules

* chore: apply suggestions from CR
2023-09-12 07:57:15 -05:00
Lei, HUANG
36231a5d50 feat(mito2): add alloc_tracker for memtable (#2266)
* feat: add alloc_tracker for memtable

* chore: integrate WriteBufferManager
2023-09-12 07:57:15 -05:00
JeremyHi
a7fa40e16d fix: filter out outdated heartbeat (#2303)
* fix: filter out outdated heartbeat, #1707

* feat: reorder handlers

* refactor: disableXXX to enableXXX

* feat: make full use of region leases to facilitate failover

* chore: minor refactor

* chore: by comment

* feat: logging on inactive/active
2023-09-12 07:57:15 -05:00
Yingwen
648b2ae293 feat(mito): Flush region (#2291)
* chore: call handle_flush_request

* feat: alias SchedulerRef and clean scheduler on drop

* feat: add scheduler to workers

* feat: remove RegionMemtableStats

* feat: pick regions to flush

* feat: add more fields to region flush task

* feat: smallvec workspace dep

* feat: Use list to hold immutable memtables

* feat: flush job wip

* feat: use access layer to read write sst

* feat: flush memtables to l0

* feat: write manifest

* feat: schedule next flush on success

* feat: schedule flush on success and failure

* feat: add purger to region

* feat: apply edit after flush

* feat: collect stats for SSTs

* feat: manual flush

* test: test flush and fix manifest test

* feat: remove flush scheduler job limit

* fix: typo

* style: clippy

* feat: clean flushed files on failure

* chore: address CR comment

* refactor: Use put_rows

* feat: Clean flush scheduler on drop

* feat: remove region flush status on drop and close

* chore: address CR comment
2023-09-12 07:57:15 -05:00
Weny Xu
fa5e3b94d3 refactor: refactor ddl procedure (#2304) 2023-09-12 07:57:15 -05:00
Weny Xu
4818887e38 refactor: refactor DistInstance (#2305) 2023-09-12 07:57:15 -05:00
Ruihang Xia
eddff17523 feat: drop region in mito2 (#2286)
* basic impl

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* check in opening region

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix typo

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix clippy

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix typo again

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* Update src/mito2/src/worker/handle_drop.rs

Co-authored-by: JeremyHi <jiachun_feng@proton.me>

* remove file in order

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix remove logic

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* use scan to list files

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: JeremyHi <jiachun_feng@proton.me>
2023-09-12 07:57:15 -05:00
Weny Xu
c839ed271c refactor: refactor: ddl context (#2301)
* refactor: refactor: ddl context

* refactor: remove unused code

* chore: apply suggestions from CR
2023-09-12 07:57:15 -05:00
JeremyHi
4d2cae4174 refactor: inactive node manager (#2300)
refactor: use region_id instead of table&region_num in InactiveNodeManager
2023-09-12 07:57:15 -05:00
Yingwen
b234733c61 feat(mito): Support deleting rows in mito2 (#2275)
* feat: check delete request

* test: test delete and overwrite
2023-09-12 07:57:15 -05:00
Lei, HUANG
9691d19601 feat: impl kv backend for raft engine (#2280)
* feat: kv backend on raft-engine

* feat: raft-engine kvbackend

* fix: toml

* fix: some review comments

* chore: optimize delete

* fix: lift lock in batch_delete
2023-09-12 07:57:15 -05:00
LFC
ff3881f0e1 feat: drop distributed Mito2 table (#2260)
* feat: drop distributed Mito2 table

* rebase develop

* fix: resolve PR comments

* fix: resolve PR comments
2023-09-12 07:57:15 -05:00
JeremyHi
fa542f6e93 feat: new heartbeat (#2299) 2023-09-12 07:57:15 -05:00
Zhenchi
d6c82867d5 refactor: remove the most Table impls (#2274)
Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
Lei, HUANG
86d56f71ef fix: flume bug (#2298)
fix: flume
2023-09-12 07:57:15 -05:00
Zhenchi
b42d343ae6 feat(frontend): unify column inserter and row inserter (#2293)
* refactor: unify inserter

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat(frontend): unify column inserter and row inserter

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: remove redundant clone

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: move empty check ahead

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* feat: add more logs

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: leading license

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* adjust indent

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-09-12 07:57:15 -05:00
Yingwen
365e557e7a feat(mito): Integrate access layer and file purger to region (#2296)
* feat: alias SchedulerRef and clean scheduler on drop

* feat: add scheduler to workers

* feat: use access layer to read write sst

* feat: add purger to region

* refactor: allow getting region_dir from AccessLayer

* feat: add scheduler to FlushScheduler

* feat: getter for object store

* chore: fix typo

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>

---------

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
Ruihang Xia
46d171d341 chore: bump greptime-proto to replace region_dir (#2290)
chore: bump greptime-proto to replace region_dir

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
Ruihang Xia
718246ea1a feat: implement heartbeat for region server (#2279)
* retrieve region stats from region server

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* implement heartbeat handler

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* start datanode with region server

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove comment

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* disable non-unit test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* implement heartbeat task

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-12 07:57:15 -05:00
JeremyHi
58d07e0e62 feat: v04 rm unused exprs (#2285)
* feat: rm compact and flush exprs

* refactor: continue to rm compact and flush
2023-09-12 07:57:15 -05:00
dennis zhuang
db89235474 feat: only allow timestamp type as time index (#2281)
* feat: only allow timestamp data type as time index

* test: update sqltest cases, todo: need some fixes

* fix: sqlness tests

* fix: forgot adding back cte test

* chore: style
2023-09-12 07:57:15 -05:00
shuiyisong
6e593401f7 refactor: collecting memory usage during scan (#2353)
* chore: try custom metrics

* chore: fix header

* chore: minor change
2023-09-12 15:52:57 +08:00
Yingwen
466fbaca5d fix: panic in try_into_vector() (#2351) 2023-09-11 19:06:31 +08:00
ZonaHe
de966af83b feat: update dashboard to v0.3.3 (#2339)
Co-authored-by: ZonaHex <ZonaHex@users.noreply.github.com>
2023-09-06 19:19:56 +08:00
Zou Wei
b8c50d00aa feat: sqlness test for interval type (#2265)
* feat: add integration-test for interval type.

* chore: add two cases.

* chore: cr

* chore: Field to Column
2023-09-04 14:30:48 +08:00
Ruihang Xia
a12ee5cab8 fix: qualify inputs on handling join in promql (#2297)
* add qualifier to join inputs

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add one more case

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update test results

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-09-01 11:51:34 +08:00
ZonaHe
a0d15b489a feat: update dashboard to v0.3.2 (#2295)
Co-authored-by: ZonaHex <ZonaHex@users.noreply.github.com>
2023-08-31 22:05:00 +08:00
shuiyisong
baa372520d fix: json compatibility to null (#2287)
* fix: existing null value for schema name value

* chore: fix null check

* fix: change catalognamevalue and schemanamevalue to option

* fix: fix null case
2023-08-31 14:21:10 +08:00
shuiyisong
5df4d44761 feat: schema level opts (#2283)
* chore: update proto

* chore: add try from for schema name value

* chore: merge schema opts to table opts while creating table

* chore: use table ttl opts first

* chore: add unit test

* chore: update proto version
2023-08-30 08:11:08 +00:00
Weny Xu
8e9f2ffce4 fix: skip procedure if target route is not found (#2277)
* fix: skip procedure if target route is not found

* chore: apply suggestions from CR
2023-08-30 06:59:50 +00:00
Weny Xu
1101e7bb18 fix: deregister table after keeper closes table (#2278)
* fix: deregister table after keeper closes table

* chore: apply suggestions from CR
2023-08-30 03:43:04 +00:00
zyy17
5fbc941023 ci: upload the latest artifacts to 'latest/' directory of S3 bucket in scheduled and formal release (#2276)
Signed-off-by: zyy17 <zyylsxm@gmail.com>
2023-08-29 09:00:45 +00:00
Bamboo1
68600a2cf9 feat(mito2): add file purger and cooperate with scheduler to purge sst files (#2251)
* feat: add file purger and use scheduler

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* feat: print some information about handling error message

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: resolve conversion

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: resolve conversation

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: resolve conflicting files

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

---------

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>
2023-08-29 07:55:03 +00:00
Yingwen
805f254d15 feat(mito): Flush framework for mito2 (#2262)
* feat: write buffer manager

* feat: skeleton

* feat: add flush logic to write path

* feat: add methods to memtable trait

* feat: freeze memtable

* feat: define flush task

* feat: schedule_flush wip

* feat: adding pending requests/tasks

* feat: separate ddl request and background request

* feat: Remove RegionTask and RequestBody

* feat: handle flush related requests

* feat: make tests pass

* style: fix clippy

* docs: update comment

* refactor: rename background requests

* feat: replace Option<RegionWriteCtx> with an enum MaybeStalling
2023-08-29 07:13:15 +00:00
Zhenchi
2a6c830ca7 refactor(table): remove Table impl for system (#2270)
* refactor(table): remove Table impl for system

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: format & import

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-29 03:43:43 +00:00
Weny Xu
22dea02485 fix: use RegionId region number instead (#2273) 2023-08-29 02:52:24 +00:00
LFC
ef75e8f7c3 feat: create distributed Mito2 table (#2246)
* feat: create distributed Mito2 table

* rebase develop
2023-08-28 12:07:52 +00:00
Weny Xu
71fc3c42d9 fix: open region does not register catalog/schema (#2271)
* fix: open region does not register catalog/schema

* fix: fix ci
2023-08-28 12:06:10 +00:00
JeremyHi
c02ac36ce8 feat: avoid confusion in desc table (#2272)
feat: Field to Column to aviod confusion in DESC TABLE
2023-08-28 11:50:33 +00:00
Lei, HUANG
c112b9a763 feat(mito2): WAL replay (#2264)
* feat: replay memtable when opening table

* test: region replay

* refactor: save logstore in TestEnv

* fix: some cr comments

* chore: rebase develop

* chore: update last entry id during replay
2023-08-28 11:45:23 +00:00
Weny Xu
96fd17aa0a fix: fix typoes (#2268) 2023-08-28 09:26:00 +00:00
Ruihang Xia
6b8cf0bbf0 feat: impl region engine for mito (#2269)
* update proto

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* convert request

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update proto

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* import result convertor

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* rename symbols

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-28 09:24:12 +00:00
Yingwen
e2522dff21 feat(mito): Skeleton for scanning a region (#2230)
* feat: define stream builder

* feat: scan region wip

* feat: create SeqScan in ScanRegion

* feat: scanner

* feat: engine handles scan request

* feat: map projection index to column id

* feat: Impl record batch stream

* refactor: change BatchConverter to ProjectionMapper

* feat: add column_ids to mapper

* feat: implement SeqScan::build()

* chore: fix typo

* docs: add mermaid for ScanRegion

* style: fix clippy

* test: fix record batch test

* fix: update sequence and entry id

* test: test query

* feat: address CR comment

* chore: address CR comments

* chore: Update src/mito2/src/read/scan_region.rs

Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>

---------

Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>
2023-08-28 06:59:31 +00:00
LFC
d8f851bef2 fix: keep region failover state not changed upon failure (#2261) 2023-08-28 04:40:47 +00:00
JeremyHi
63b22b2403 feat: prometheus row inserter (#2263)
* feat: prometheus row inserter

* chore: add unit test

* refactor: to row_insert_requests

* chore: typo

* chore: alloc row by TableData

* chore: by review comment
2023-08-28 03:22:23 +00:00
Weny Xu
c56f5e39cd refactor: set default metasrv procedure retry times to 12 (#2242) 2023-08-26 07:41:15 +00:00
Weny Xu
7ff200c0fa fix: align region numbers to real regions (#2257) 2023-08-25 11:48:58 +00:00
dennis zhuang
5160838d04 chore: change version to 0.4.0-nightly (#2258)
* chore: change version to 0.4.0-nightly

* fix: test
2023-08-25 09:44:39 +00:00
shuiyisong
f16f58266e refactor: query_ctx from http middleware (#2253)
* chore: change userinfo to query_ctx in http handler

* chore: minor change

* chore: move prometheus http to http mod

* chore: fix uni test:

* chore: add back schema check

* chore: minor change

* chore: remove clone
2023-08-25 09:36:33 +00:00
Ruihang Xia
8d446ed741 fix: quote ident on rendered SQL (#2248)
* fix: quote ident on rendered SQL

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* read quote style from query context

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update sqlness result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-25 07:25:21 +00:00
JeremyHi
de1daec680 feat: upgrade desc table output (#2256) 2023-08-25 06:52:22 +00:00
Zhenchi
9d87c8b6de refactor(table): cleanup dist table (#2255)
Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-25 06:37:39 +00:00
Lei, HUANG
6bf260a05c chore: write to mito2 (#2250)
* chore: write to mito2

* fix: clippy

* feat: brdige memtable

* chore: rebase develop
2023-08-25 06:18:42 +00:00
WU Jingdi
15912afd96 fix: the inconsistent order of input/output in range select (#2229)
* fix: the inconsistent order of input/output in range select

* chore: apply CR
2023-08-25 04:12:59 +00:00
Lei, HUANG
dbe0e95f2f feat(mito2): concat and projection (#2243)
* refactor: use arrow::compute::concat instead of push values to vector builders

* feat: support projection

* refactor: remove sequence

* refactor: concatenate

* fix: series must not be empty

* refactor: projection
2023-08-25 03:25:27 +00:00
Ruihang Xia
20b7f907b2 fix: promql planner should clear its states on each selector (#2247)
* reset planner status on selector

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add sqlness test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add empty line

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* sort result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* mask fields to keep ordering

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-25 03:07:44 +00:00
Weny Xu
b13d932e4e fix: fix RegionAliveKeeper does not find the table after restarting (#2249) 2023-08-25 03:05:17 +00:00
Bamboo1
48348aa364 fix: fix test_scheduler_continuous_stop in scheduler (#2252)
* fix: fix test_scheduler_continuous_stop in scheduler

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: add document annotation

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

---------

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>
2023-08-25 02:59:48 +00:00
Zhenchi
9ce73e7ca1 refactor(frontend): TableScan instead of scan_to_stream for COPY TO (#2244)
* refactor(frontend): TableScan instead of `scan_to_stream` for `COPY TO`

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: format

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-24 12:46:54 +00:00
Ruihang Xia
b633a16667 feat: apply rewriter to subquery exprs (#2245)
* apply rewriter to subquery exprs

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* workaround for datafusion's check

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean up

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add sqlness test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix typo

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* change time index type

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-24 11:48:04 +00:00
Zhenchi
0a6ab2a287 refactor(script): not to call scan_to_stream on table (#2241)
* refactor(script): not to call `scan_to_stream` on table

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: build plan via LogicalPlanBuilder

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-24 08:10:07 +00:00
JeremyHi
7746e5b172 feat: dist row inserter (#2231)
* feat: fronend row inserter

* feat: row splitter

chore: row splitter's unit test

* feat: RowDistInserter

* feat: make influxdb line protocol using row-based protocol

* Update src/partition/src/row_splitter.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

* Update src/frontend/src/instance/distributed/row_inserter.rs

Co-authored-by: Yingwen <realevenyag@gmail.com>

* chore: by review comment

* Update src/frontend/src/instance/distributed/row_inserter.rs

Co-authored-by: LFC <bayinamine@gmail.com>

* chore: by comment

---------

Co-authored-by: Yingwen <realevenyag@gmail.com>
Co-authored-by: LFC <bayinamine@gmail.com>
2023-08-24 06:58:05 +00:00
Weny Xu
a7e0e2330e fix: invalidate cache after altering (#2239) 2023-08-24 03:56:17 +00:00
Lei, HUANG
19d2d77b41 fix: parse large timestamp (#2185)
* feat: support parsing large timestamp values

* chore: update sqlness tests

* fix: tests

* fix: allow larger window
2023-08-24 03:52:15 +00:00
Yingwen
4ee1034012 feat(mito): merge reader for mito2 (#2210)
* feat: Implement slice and first/last timestamp for Batch

* feat(mito): implements sort/concat for Batch

* chore: fix typo

* chore: remove comments

* feat: sort and dedup

* test: test batch operations

* chore: cast enum to test op type

* test: test filter related api

* sytle: fix clippy

* feat: implement Node and CompareFirst

* feat: merge reader wip

* feat: merge wip

* feat: use batch's operation to sort and dedup

* feat: implement BatchReader for MergeReader

* feat: simplify codes

* test: test merge reader

* refactor: use test util to create batch

* refactor: remove unused imports

* feat: update comment

* chore: remove metadata() from Source

* chroe: update comment

* feat: source supports batch iterator

* chore: update comment
2023-08-24 03:37:51 +00:00
Ruihang Xia
e5ba3d1708 feat: rewrite the dist analyzer (#2238)
* it works!

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* clean up

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add documents

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove unstable timestamp from sqlness test

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* rename rewriter struct

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-24 03:29:08 +00:00
dennis zhuang
8b1f4eb958 feat: types sqlness tests (#2073)
* feat: timestamp types sqlness tests

* feat: adds timestamp tests

* test: add string tests

* test: comment a case in timestamp

* test: add float type tests

* chore: adds TODO

* feat: set TZ=UTC for sqlness test
2023-08-24 03:26:19 +00:00
discord9
eca7e87129 chore: try from value (#2236)
* chore: try from value

* chore: add TryFromValueError variant
2023-08-24 02:44:13 +00:00
Weny Xu
beb92ba1d2 refactor: use table id instead of table ident (#2233) 2023-08-23 13:28:08 +00:00
Lei, HUANG
fdb5ad23bf refactor: use Batch::sort_and_dedup instead of Values::sort_in_place (#2235) 2023-08-23 08:56:49 +00:00
Ruihang Xia
d581688fd2 fix: dist planner has wrong behavior in table with multiple partitions (#2237)
* fix: dist planner has wrong behavior in table with multiple partitions

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* Update tests/cases/distributed/explain/multi_partitions.sql

Co-authored-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-23 08:32:20 +00:00
Bamboo1
4dbc32f532 refactor: remove associate type in scheduler to simplify it #2153 (#2194)
* feature: add a simple scheduler using flume

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: only use a sender rather clone many senders

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: use select to avoid loop

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* feat: add parameters in new function to build the flume capacity and number of receivers

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* test: add countdownlatch test concurrency

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* test: add barrier replacing countdownlatch to test concurrency and add wait all tasks finished in stop

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: add some document annotation

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: add license header

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: add Cargo.lock

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: Cargo.toml format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: delete println in test

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* feat: add error handle

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: fix error handle and add test scheduler stop

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: spelling mistake

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: wait all tasks finished

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: add todo which need wrap Future returned by send_async

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* test: remove unnessary sleep in test

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: resolve some conflicts

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* fix: resolve conversation

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* chore: code format

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

* feat: modify the function of schedule to synchronize and drop sender after stopping scheduler

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>

---------

Signed-off-by: ZhuZiyi <zyzhu2001@gmail.com>
2023-08-23 06:28:00 +00:00
Zhenchi
af95e46512 refactor(table): eliminate calls to DistTable.delete (#2225)
* refactor(table): eliminate calls to DistTable.delete

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: format

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: clippy

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-23 02:33:48 +00:00
Weny Xu
d81ddd8879 chore: fix clippy (#2232) 2023-08-23 02:24:29 +00:00
Ning Sun
88247e4284 fix!: resolve residual issues with removing prometheus port (#2227)
* fix: resolve residual issues when removing prometheus port

* fix: remove prometheus from sample config as well
2023-08-23 01:49:11 +00:00
Ruihang Xia
18250c4803 feat: implement Flight and gRPC services for RegionServer (#2226)
* extract FlightCraft trait

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* split service handler in GrpcServer

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* left grpc server implement

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* start region server if configured

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-22 13:30:09 +00:00
dennis zhuang
18fa0e01ed feat: remove checkpoint_on_startup (#2228)
feat: update flushed manifest version when it is larger
2023-08-22 13:09:34 +00:00
Yingwen
cc3e198975 feat(mito): Implement operations like concat and sort for Batch (#2203)
* feat: Implement slice and first/last timestamp for Batch

* feat(mito): implements sort/concat for Batch

* chore: fix typo

* chore: remove comments

* feat: sort and dedup

* test: test batch operations

* chore: cast enum to test op type

* test: test filter related api

* sytle: fix clippy

* docs: comment for slice

* chore: address CR comment

Don't return Option in get_timestamp()/get_sequence()
2023-08-22 12:03:02 +00:00
Yingwen
cd3755c615 feat(mito): Support handling RegionWriteRequest (#2218)
* feat: convert region request to worker write request

* chore: remove unused codes

* test: fix tests compiler errors

* chore: remove create/close/open request from worker requests

* chore: add comment

* chore: fix typo
2023-08-22 11:16:00 +00:00
Lei, HUANG
be1e13c713 feat(mito2): time series memtable (#2208)
* feat: time series memtable

* feat: add some test

* fix: some clippy warnings

* chore: some rustdoc

* refactor: test

* fix: remove useless functions

* feat: add config for TimeSeriesMemtable

* chore: some optimize

* refactor: remove bucketing

* refactor: avoid cloing RegionMetadataRef across all Series; make initial_builder_capacity a const; sort batch only by timestamp and sequence
2023-08-22 08:40:46 +00:00
Zhenchi
cb3561f3b3 refactor(table): eliminate calls to DistTable.insert (#2219)
Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-22 06:15:02 +00:00
Niwaka
b3b43fe1c3 fix: table options can't be found in distributed mode (#2209)
* fix: table options can't be found in distributed mode

* refactor: use iterator for regions_numbers

* chore: remove TODO
2023-08-22 03:53:56 +00:00
WU Jingdi
b411769de6 feat: Implement a basical range select query (#2138)
* feat: Implement a basical range select query

* chore: support any timestamp type & CR fix
2023-08-22 03:07:14 +00:00
niebayes
e5f4ca2dab feat: streaming do_get (#2171)
* feat: rewrite do_get for streaming get flight data

* feat: rewrite do_get call stack but leave the async stream adapter not modified yet

* feat: rewrite the async stream adapter to accept greptime record batch stream

* fix: resolve some PR comments

* feat: rewrite tests to adapt to the streaming do_get

* feat: add unit tests for streaming do_get

* feat: rewrite timer metric of merge scan

* remove unhelpful unit tests for streaming do_get

* add a new metric timer for merge scan and fix some test errors

* rewrite mysql writer to write query results in a streaming manner

* fix: fix fmt errors

* fix: rewrite sqlness runner to take into account the streaming do_get

* fix: fix toml format errors

* fix: resolve some PR comments

* fix: resolve some PR comments

* fix: refactor do_get to increase readability

* fix: refactor mysql try_write_one to increase readability
2023-08-22 02:54:05 +00:00
Weny Xu
5b7b2cf77d fix: fix ddl client can not update leader addr (#2205)
* fix: fix ddl client can not update leader addr

* chore: apply suggestions from CR

* feat: add message to context

* fix: only retry if unavailable or deadline exceeded

* chore: apply suggestions from CR
2023-08-21 13:57:29 +00:00
shuiyisong
9352649f22 chore: add table region key to delete in upgrade tool (#2214) 2023-08-21 08:16:10 +00:00
shuiyisong
c5f507c20e fix: add user_info extension to prom_store handler (#2212)
chore: add user_info extention to prom_store auth
2023-08-21 04:55:34 +00:00
JeremyHi
033b650d0d feat: row write protocol (#2189)
* feat: datanode's row insrter

* refactor: ExprFactory

* feat: row inserter in standalon mode

* chore: minor refactor

* feat: influxdb line protocol's row protocol

* chore: minor refactor

* improve: avoid to use too many string

* no longer async

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>

* chore: do not check empty data

* chore: by review comment

* chore: by comment

* chore: by review comment

* chore: by review comment

---------

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-19 13:08:44 +00:00
dennis zhuang
272f649b22 fix: some TODO in sqlness cases and refactor meta-client error (#2207)
* fix: some TODO in sqlness cases and refactor meta-client error

* fix: delete tests/cases/standalone/alter/drop_col_not_null_next.output
2023-08-18 10:09:11 +00:00
Ruihang Xia
3150f4b22e fix: specify input ordering and distribution for prom plan (#2204)
* fix: specify input ordering and distribution for prom plan

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update sqlness result

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-18 09:45:46 +00:00
Weny Xu
e1ce1d86a1 refactor: unite key serialization method (#2195) 2023-08-18 09:42:19 +00:00
ZonaHe
b8595e1960 feat: update dashboard to v0.3.1 (#2192)
Co-authored-by: ZonaHex <ZonaHex@users.noreply.github.com>
2023-08-18 09:42:18 +00:00
shuiyisong
61e6656fea fix: auth in prometheus gateway service (#2206)
* fix: auth in prometheus gateway service

* chore: remove unused code

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>

---------

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-18 09:41:38 +00:00
Ruihang Xia
1bbec75f5b fix: skip partition clause in show create table (#2200)
* fix: skip partition clause in show create table

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update test results

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-18 09:10:31 +00:00
Zhenchi
8d6a2d0b59 refactor: apply numbers to ThinTable (#2202)
* refactor: apply numbers to `ThinTable`

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: tiny polish

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: unused import

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-18 03:11:37 +00:00
Weny Xu
177036475a fix: support to copy from parquet with typecast (#2201) 2023-08-18 03:09:54 +00:00
Zhenchi
87a730658a refactor: add ThinTable to proxy tables from infoschema (#2193)
* refactor: add thin table to proxy tables in info_schema

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix(catalog): fix typo in DataSourceAdapter struct name

* fix: remove redundant Send + Sync

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor(catalog): rename DataSourceAdapter to InformationTableDataSource

* feat(catalog): add ThinTableAdapter for adapting ThinTable to Table interface

* rebase develop

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: default impl for table_type of InformationTable

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* refactor: filter_pushdown as table field

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* fix: remove explicit type declaration

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-17 15:19:14 +00:00
JeremyHi
b67e5bbf70 fix: invalid err msg (#2196) 2023-08-17 11:12:35 +00:00
Ruihang Xia
4aaf6aa51b feat: implement query API for RegionServer (#2197)
* some initial change

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* impl dummy structs

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* decode and send logical plan

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* implement table scan

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* add some comments

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-17 11:02:31 +00:00
Weny Xu
6e6ff5a606 refactor: update table metadata in single txn (#2172)
* refactor: table-metadata-manager

* feat: remove comparing when deleting metadata

* fix: fix comment typos

* chore: apply suggestions from CR

* test: add tests for updating DatanodeTable

* fix: fix clippy

* chore: apply suggestions from CR

* refactor: improve update table route tests

* refactor: return Txn instead of TxnRequest

* chore: apply suggestions from CR

* chore: apply suggestions from CR

* refactor: update table metadata in single txn

* feat: check table exists before drop table executing

* test: add tests for table metadata manager

* refactor: remove table region manager

* chore: apply suggestions from CR

* feat: add bench program

* chore: apply suggestions from CR
2023-08-17 06:29:19 +00:00
Yingwen
4ba12155fe feat(mito): Implement SST format for mito2 (#2178)
* chore: update comment

* feat: stream writer takes arrow's types

* feat: Define Batch struct

* feat: arrow_schema_to_store

* refactor: rename

* feat: write parquet in new format with tsids

* feat: reader support projection

* feat: Impl read compat

* refactor: rename SchemaCompat to CompatRecordBatch

* feat: changing sst format

* feat: make it compile

* feat: remove tsid and some structs

* feat: from_sst_record_batch wip

* chore: push array

* chore: wip

* feat: decode batches from RecordBatch

* feat: reader converts record batches

* feat: remove compat mod

* chore: remove some codes

* feat: sort fields by column id

* test: test to_sst_arrow_schema

* feat: do not sort fields

* test: more test helpers

* feat: simplify projection

* fix: projection indices is incorrect

* refactor: define write/read format

* test: test write format

* test: test projection

* test: test convert record batch

* feat: remove unused errors

* refactor: wrap get_field_batch_columns

* chore: clippy

* chore: fix clippy

* feat: build arrow schema from region meta in ReadFormat

* feat: initialize the parquet reader at `build()`

* chore: fix typo
2023-08-17 06:25:50 +00:00
Weny Xu
832e5dcfd7 chore: remove allow-unused (#2184) 2023-08-17 03:15:12 +00:00
shuiyisong
d45ee8b42a chore: fix collect region stat on non-base table (#2190) 2023-08-17 02:13:49 +00:00
JeremyHi
6cd7319d67 refactor: grpc insert (#2188)
* feat: interval type for row protocol

* feat: minor refactor grpc insert

* Update src/common/grpc-expr/src/util.rs

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>

* fix: by comment

---------

Co-authored-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-16 11:25:25 +00:00
Yingwen
bb062003ef ci: fallback to run_id to avoid cancelling other jobs (#2186)
ci: fallback to run id to avoid cancelling other jobs
2023-08-16 09:24:17 +00:00
Weny Xu
8ea1763033 refactor: refactor table metadata manager (#2159)
* refactor: table-metadata-manager

* feat: remove comparing when deleting metadata

* fix: fix comment typos

* chore: apply suggestions from CR

* test: add tests for updating DatanodeTable

* fix: fix clippy

* chore: apply suggestions from CR

* refactor: improve update table route tests

* refactor: return Txn instead of TxnRequest

* chore: apply suggestions from CR

* chore: apply suggestions from CR
2023-08-16 06:43:03 +00:00
Zhenchi
1afe96e397 refactor: prevent dist table from invoking scan (#2179)
* refactor: prevent dist table from invoking `scan`

* refactor: reorg code

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

* chore: add comment

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2023-08-16 04:43:33 +00:00
Ruihang Xia
814c599029 ci: cancel in-progress actions on new commit (#2182)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-16 04:21:14 +00:00
Ruihang Xia
4c3169431b feat: move region metadata to store-api (#2181)
* add metadata & handle_read

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* move metadata to store-api

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* dep aquamarine

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove deadcode

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove temporary code

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* Update src/store-api/Cargo.toml

Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>

* remove old mod

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>
2023-08-16 04:18:26 +00:00
sh2
202540823f refactor!: move prometheus routes to default http server (#2005)
* move prometheus routes to default http server

Signed-off-by: sh2 <shawnhxh@outlook.com>

* fix ci test and remove the server logic of prometheus

* remove unused import and prometheus relevant code

* fix ci: rustfmt and test

* fix ci: silly fmt

* fix ci: silly silly fmt

* change `/prom_store` back to `/prometheus`

* remove unused variable

---------

Signed-off-by: sh2 <shawnhxh@outlook.com>
2023-08-16 03:21:14 +00:00
dennis zhuang
0967678a51 feat: don't enable telemetry for debug building (#2177) 2023-08-16 01:53:11 +00:00
shuiyisong
c8cde704cf chore: minor auth crate change (#2176)
* chore: pub auth_mysql

* chore: pub all error

* chore: remove back to error

* chore: wrap failed permission check result to err

* chore: minor change
2023-08-15 10:49:22 +00:00
JeremyHi
24dc827ff9 feat: grpc handler result (#2107)
* feat: grpc handler inner result

* feat: ext header, x-greptime-err-code, x-greptime-err-msg

* fix: sqlness case

* chore: by comment

* fix: convert status to Error
2023-08-15 09:34:00 +00:00
Weny Xu
f5e44ba4cf docs: rfc of update metadata in single txn (#2165)
* docs: rfc of update metadata in single txn

* chore: apply suggestion from CR
2023-08-15 17:44:07 +08:00
zyy17
32c3ac4fcf refactor: improve the image building performance (#2175)
* refactor: use '--output type=local' in 'build-greptime-by-buildx' target to reduce unnecessary 'docker cp'"

Signed-off-by: zyy17 <zyylsxm@gmail.com>

* refactor: improve the image building performance

* ci: release centos dev builder

* ci: use 'make build-by-dev-builder' to improve docker build performance

* refactor: add 'which' command in centos

* fix: add 'OUTPUT_DIR' to fix 'make docker-image-buildx' error

* fix: fix incorrect dockerfile path

Signed-off-by: zyy17 <zyylsxm@gmail.com>

* refactor: remove configure-aws-credentials action and use env variables

Signed-off-by: zyy17 <zyylsxm@gmail.com>

* ci: update slack notification prompt

* refactor: clean up the target directory before building artifacts of centos7

---------

Signed-off-by: zyy17 <zyylsxm@gmail.com>
2023-08-15 09:28:09 +00:00
Niwaka
a8f2e4468d feat: handle multiple grpc deletes (#2150)
* feat: handle multiple grpc deletes

* fix: make DistDeleter::grpc_delete return usize

* fix: remove backtrace from MissingTimeIndexColumn

* fix: avoid using unwrap in PartitionRuleManager::split_delete_request

* fix: simplify MissingTimeIndexColumn
2023-08-15 08:22:46 +00:00
Yingwen
d4565c0a94 feat(mito): Defines the read Batch struct for mito2 (#2174)
* feat: define batch

* feat: define Batch struct

* feat: stream writer takes arrow's types

* feat: make it compile

* feat: use uint64vector and uint8vector

* feat: add timestamps and primary key
2023-08-15 06:39:21 +00:00
Ruihang Xia
2168970814 feat: define region server and related requests (#2160)
* define region server and related requests

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fill request body

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* change mito2's request type

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix clippy

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* chore: bump greptime-proto to d9167cab (row insert/delete)

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix test compile

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* remove name_to_index

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* address cr comments

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* finilise

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-15 06:27:27 +00:00
Weny Xu
69a2036cee feat!: add deserializer for Partition (#2169)
* feat!: add deserializer for Partition

* fix: fix tests
2023-08-15 03:36:58 +00:00
Lei, HUANG
e924b44e83 refactor: KeyValues return ValueRef (#2170)
* refactor: KeyValues return ValueRef

* 1. Change KeyValues returned value from pb value to ValueRef
2. Replace OpType/SemanticType with pb's OpType and SemanticType to avoid duplicated conversions.

* feat: define min value of OpType as a const

* fix: toml format
2023-08-14 14:51:13 +00:00
Yingwen
768239eb49 fix: panic on truncate table in distributed mode (#2173) 2023-08-14 14:20:20 +00:00
Ning Sun
f3157df190 fix: normalize otlp string keys (#2168) 2023-08-14 09:39:54 +00:00
dennis zhuang
b353bd20db fix: print_anonymous_usage_data_disclaimer at wrong place (#2167) 2023-08-14 08:01:10 +00:00
Lei, HUANG
55b5df9c51 feat: row wise converter (#2162)
* feat: impl mem-comparable encoding for timestamp

* fix: test cases

* impl time series encode/decoder

* fix: merge unsupported match arms

* fix: clippy

* chore: big number delimiter

* feat: encode timestamps as i64

* fix: remove useless error variant
2023-08-14 07:13:39 +00:00
Ruihang Xia
393047a541 feat: implement metric for MergeScanExec (#2166)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-14 07:10:45 +00:00
LFC
606b489d53 feat: redact secrets in sql when logging (#2141) 2023-08-14 06:40:00 +00:00
Weny Xu
d0b3607633 feat: add table route manager and upgrade tool (#2145)
* feat: add table route manager and upgrade tool

* test: add table route manager tests

* feat: add new TableRouteValue struct

* chore: apply suggestions from CR

* refactor: change HashMap to BTreeMap

* feat: add version to TableRouteValue
2023-08-14 04:19:44 +00:00
Weny Xu
5b012a1f67 feat!: switch to new catalog/schema key (#2140)
* feat!: switch to new catalog/schema key

* chore: apply suggestions from CR
2023-08-14 03:08:43 +00:00
Ruihang Xia
f6b53984da fix(metasrv)!: do not overwrite boolean options unconditionally (#2161)
* fix: do not overwrite boolean options unconditionally

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* fix sqlness start command

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2023-08-14 03:04:54 +00:00
shuiyisong
7f51141ed0 refactor: auth crate (#2148)
* chore: move user_info to auth crate

* chore: temp commit before resolving tests compile error

* chore: fix compile issue

* chore: minor fix

* chore: tmp save

* chore: change user_info to trait

* chore: minor change & use auth result user info in pg session setup

* chore: add as_any to user_info

* chore: rename user_info

* chore: remove ice file

* chore: add permission checker

* chore: add grpc permission check

* chore: add session spawn user_info to query_ctx

* chore: minor update

* chore: add permission checker to sql handler & temp save

* chore: add permission checker to prometheus handler

* chore: add permission checker to opentsdb handler

* chore: add permission checker to other handlers

* chore: add test

* chore: add user_info setting on http entrance

* chore: fix toml

* chore: remove box in permission req

* chore: cr issue

* chore: cr issue
2023-08-14 02:51:26 +00:00
996 changed files with 62031 additions and 41884 deletions

View File

@@ -1,70 +0,0 @@
name: Build and push dev-builder image
description: Build and push dev-builder image to DockerHub and ACR
inputs:
dockerhub-image-registry:
description: The dockerhub image registry to store the images
required: false
default: docker.io
dockerhub-image-registry-username:
description: The dockerhub username to login to the image registry
required: true
dockerhub-image-registry-token:
description: The dockerhub token to login to the image registry
required: true
dockerhub-image-namespace:
description: The dockerhub namespace of the image registry to store the images
required: false
default: greptime
acr-image-registry:
description: The ACR image registry to store the images
required: true
acr-image-registry-username:
description: The ACR username to login to the image registry
required: true
acr-image-registry-password:
description: The ACR password to login to the image registry
required: true
acr-image-namespace:
description: The ACR namespace of the image registry to store the images
required: false
default: greptime
version:
description: Version of the dev-builder
required: false
default: latest
runs:
using: composite
steps:
- name: Login to Dockerhub
uses: docker/login-action@v2
with:
registry: ${{ inputs.dockerhub-image-registry }}
username: ${{ inputs.dockerhub-image-registry-username }}
password: ${{ inputs.dockerhub-image-registry-token }}
- name: Build and push dev builder image to dockerhub
shell: bash
run:
make dev-builder \
BUILDX_MULTI_PLATFORM_BUILD=true \
IMAGE_REGISTRY=${{ inputs.dockerhub-image-registry }} \
IMAGE_NAMESPACE=${{ inputs.dockerhub-image-namespace }} \
IMAGE_TAG=${{ inputs.version }}
- name: Login to ACR
uses: docker/login-action@v2
continue-on-error: true
with:
registry: ${{ inputs.acr-image-registry }}
username: ${{ inputs.acr-image-registry-username }}
password: ${{ inputs.acr-image-registry-password }}
- name: Build and push dev builder image to ACR
shell: bash
continue-on-error: true
run: # buildx will cache the images that already built, so it will not take long time to build the images again.
make dev-builder \
BUILDX_MULTI_PLATFORM_BUILD=true \
IMAGE_REGISTRY=${{ inputs.acr-image-registry }} \
IMAGE_NAMESPACE=${{ inputs.acr-image-namespace }} \
IMAGE_TAG=${{ inputs.version }}

View File

@@ -0,0 +1,76 @@
name: Build and push dev-builder images
description: Build and push dev-builder images to DockerHub and ACR
inputs:
dockerhub-image-registry:
description: The dockerhub image registry to store the images
required: false
default: docker.io
dockerhub-image-registry-username:
description: The dockerhub username to login to the image registry
required: true
dockerhub-image-registry-token:
description: The dockerhub token to login to the image registry
required: true
dockerhub-image-namespace:
description: The dockerhub namespace of the image registry to store the images
required: false
default: greptime
version:
description: Version of the dev-builder
required: false
default: latest
build-dev-builder-ubuntu:
description: Build dev-builder-ubuntu image
required: false
default: 'true'
build-dev-builder-centos:
description: Build dev-builder-centos image
required: false
default: 'true'
build-dev-builder-android:
description: Build dev-builder-android image
required: false
default: 'true'
runs:
using: composite
steps:
- name: Login to Dockerhub
uses: docker/login-action@v2
with:
registry: ${{ inputs.dockerhub-image-registry }}
username: ${{ inputs.dockerhub-image-registry-username }}
password: ${{ inputs.dockerhub-image-registry-token }}
- name: Build and push dev-builder-ubuntu image
shell: bash
if: ${{ inputs.build-dev-builder-ubuntu == 'true' }}
run: |
make dev-builder \
BASE_IMAGE=ubuntu \
BUILDX_MULTI_PLATFORM_BUILD=true \
IMAGE_REGISTRY=${{ inputs.dockerhub-image-registry }} \
IMAGE_NAMESPACE=${{ inputs.dockerhub-image-namespace }} \
IMAGE_TAG=${{ inputs.version }}
- name: Build and push dev-builder-centos image
shell: bash
if: ${{ inputs.build-dev-builder-centos == 'true' }}
run: |
make dev-builder \
BASE_IMAGE=centos \
BUILDX_MULTI_PLATFORM_BUILD=true \
IMAGE_REGISTRY=${{ inputs.dockerhub-image-registry }} \
IMAGE_NAMESPACE=${{ inputs.dockerhub-image-namespace }} \
IMAGE_TAG=${{ inputs.version }}
- name: Build and push dev-builder-android image # Only build image for amd64 platform.
shell: bash
if: ${{ inputs.build-dev-builder-android == 'true' }}
run: |
make dev-builder \
BASE_IMAGE=android \
IMAGE_REGISTRY=${{ inputs.dockerhub-image-registry }} \
IMAGE_NAMESPACE=${{ inputs.dockerhub-image-namespace }} \
IMAGE_TAG=${{ inputs.version }} && \
docker push ${{ inputs.dockerhub-image-registry }}/${{ inputs.dockerhub-image-namespace }}/dev-builder-android:${{ inputs.version }}

View File

@@ -16,47 +16,48 @@ inputs:
version:
description: Version of the artifact
required: true
release-to-s3-bucket:
description: S3 bucket to store released artifacts
required: true
aws-access-key-id:
description: AWS access key id
required: true
aws-secret-access-key:
description: AWS secret access key
required: true
aws-region:
description: AWS region
required: true
upload-to-s3:
description: Upload to S3
required: false
default: 'true'
working-dir:
description: Working directory to build the artifacts
required: false
default: .
build-android-artifacts:
description: Build android artifacts
required: false
default: 'false'
runs:
using: composite
steps:
- name: Build greptime binary
shell: bash
if: ${{ inputs.build-android-artifacts == 'false' }}
run: |
cd ${{ inputs.working-dir }} && \
make build-greptime-by-buildx \
make build-by-dev-builder \
CARGO_PROFILE=${{ inputs.cargo-profile }} \
FEATURES=${{ inputs.features }} \
BASE_IMAGE=${{ inputs.base-image }}
- name: Upload artifacts
uses: ./.github/actions/upload-artifacts
if: ${{ inputs.build-android-artifacts == 'false' }}
with:
artifacts-dir: ${{ inputs.artifacts-dir }}
target-file: ./greptime
target-file: ./target/${{ inputs.cargo-profile }}/greptime
version: ${{ inputs.version }}
working-dir: ${{ inputs.working-dir }}
# TODO(zyy17): We can remove build-android-artifacts flag in the future.
- name: Build greptime binary
shell: bash
if: ${{ inputs.build-android-artifacts == 'true' }}
run: |
cd ${{ inputs.working-dir }} && make strip-android-bin
- name: Upload android artifacts
uses: ./.github/actions/upload-artifacts
if: ${{ inputs.build-android-artifacts == 'true' }}
with:
artifacts-dir: ${{ inputs.artifacts-dir }}
target-file: ./target/aarch64-linux-android/release/greptime
version: ${{ inputs.version }}
release-to-s3-bucket: ${{ inputs.release-to-s3-bucket }}
aws-access-key-id: ${{ inputs.aws-access-key-id }}
aws-secret-access-key: ${{ inputs.aws-secret-access-key }}
aws-region: ${{ inputs.aws-region }}
upload-to-s3: ${{ inputs.upload-to-s3 }}
working-dir: ${{ inputs.working-dir }}

View File

@@ -40,7 +40,7 @@ runs:
image-registry-password: ${{ inputs.image-registry-password }}
image-name: ${{ inputs.image-name }}
image-tag: ${{ inputs.version }}
docker-file: docker/ci/Dockerfile
docker-file: docker/ci/ubuntu/Dockerfile
amd64-artifact-name: greptime-linux-amd64-pyo3-${{ inputs.version }}
arm64-artifact-name: greptime-linux-arm64-pyo3-${{ inputs.version }}
platforms: linux/amd64,linux/arm64
@@ -56,7 +56,7 @@ runs:
image-registry-password: ${{ inputs.image-registry-password }}
image-name: ${{ inputs.image-name }}-centos
image-tag: ${{ inputs.version }}
docker-file: docker/ci/Dockerfile-centos
docker-file: docker/ci/centos/Dockerfile
amd64-artifact-name: greptime-linux-amd64-centos-${{ inputs.version }}
platforms: linux/amd64
push-latest-tag: ${{ inputs.push-latest-tag }}

View File

@@ -13,26 +13,10 @@ inputs:
disable-run-tests:
description: Disable running integration tests
required: true
release-to-s3-bucket:
description: S3 bucket to store released artifacts
required: true
aws-access-key-id:
description: AWS access key id
required: true
aws-secret-access-key:
description: AWS secret access key
required: true
aws-region:
description: AWS region
required: true
dev-mode:
description: Enable dev mode, only build standard greptime
required: false
default: 'false'
upload-to-s3:
description: Upload to S3
required: false
default: 'true'
working-dir:
description: Working directory to build the artifacts
required: false
@@ -64,11 +48,6 @@ runs:
cargo-profile: ${{ inputs.cargo-profile }}
artifacts-dir: greptime-linux-${{ inputs.arch }}-pyo3-${{ inputs.version }}
version: ${{ inputs.version }}
release-to-s3-bucket: ${{ inputs.release-to-s3-bucket }}
aws-access-key-id: ${{ inputs.aws-access-key-id }}
aws-secret-access-key: ${{ inputs.aws-secret-access-key }}
aws-region: ${{ inputs.aws-region }}
upload-to-s3: ${{ inputs.upload-to-s3 }}
working-dir: ${{ inputs.working-dir }}
- name: Build greptime without pyo3
@@ -80,13 +59,13 @@ runs:
cargo-profile: ${{ inputs.cargo-profile }}
artifacts-dir: greptime-linux-${{ inputs.arch }}-${{ inputs.version }}
version: ${{ inputs.version }}
release-to-s3-bucket: ${{ inputs.release-to-s3-bucket }}
aws-access-key-id: ${{ inputs.aws-access-key-id }}
aws-secret-access-key: ${{ inputs.aws-secret-access-key }}
aws-region: ${{ inputs.aws-region }}
upload-to-s3: ${{ inputs.upload-to-s3 }}
working-dir: ${{ inputs.working-dir }}
- name: Clean up the target directory # Clean up the target directory for the centos7 base image, or it will still use the objects of last build.
shell: bash
run: |
rm -rf ./target/
- name: Build greptime on centos base image
uses: ./.github/actions/build-greptime-binary
if: ${{ inputs.arch == 'amd64' && inputs.dev-mode == 'false' }} # Only build centos7 base image for amd64.
@@ -96,9 +75,14 @@ runs:
cargo-profile: ${{ inputs.cargo-profile }}
artifacts-dir: greptime-linux-${{ inputs.arch }}-centos-${{ inputs.version }}
version: ${{ inputs.version }}
release-to-s3-bucket: ${{ inputs.release-to-s3-bucket }}
aws-access-key-id: ${{ inputs.aws-access-key-id }}
aws-secret-access-key: ${{ inputs.aws-secret-access-key }}
aws-region: ${{ inputs.aws-region }}
upload-to-s3: ${{ inputs.upload-to-s3 }}
working-dir: ${{ inputs.working-dir }}
- name: Build greptime on android base image
uses: ./.github/actions/build-greptime-binary
if: ${{ inputs.arch == 'amd64' && inputs.dev-mode == 'false' }} # Only build android base image on amd64.
with:
base-image: android
artifacts-dir: greptime-android-arm64-${{ inputs.version }}
version: ${{ inputs.version }}
working-dir: ${{ inputs.working-dir }}
build-android-artifacts: true

View File

@@ -19,21 +19,9 @@ inputs:
disable-run-tests:
description: Disable running integration tests
required: true
release-to-s3-bucket:
description: S3 bucket to store released artifacts
required: true
artifacts-dir:
description: Directory to store artifacts
required: true
aws-access-key-id:
description: AWS access key id
required: true
aws-secret-access-key:
description: AWS secret access key
required: true
aws-region:
description: AWS region
required: true
runs:
using: composite
steps:
@@ -99,7 +87,3 @@ runs:
artifacts-dir: ${{ inputs.artifacts-dir }}
target-file: target/${{ inputs.arch }}/${{ inputs.cargo-profile }}/greptime
version: ${{ inputs.version }}
release-to-s3-bucket: ${{ inputs.release-to-s3-bucket }}
aws-access-key-id: ${{ inputs.aws-access-key-id }}
aws-secret-access-key: ${{ inputs.aws-secret-access-key }}
aws-region: ${{ inputs.aws-region }}

View File

@@ -0,0 +1,80 @@
name: Build Windows artifacts
description: Build Windows artifacts
inputs:
arch:
description: Architecture to build
required: true
rust-toolchain:
description: Rust toolchain to use
required: true
cargo-profile:
description: Cargo profile to build
required: true
features:
description: Cargo features to build
required: true
version:
description: Version of the artifact
required: true
disable-run-tests:
description: Disable running integration tests
required: true
artifacts-dir:
description: Directory to store artifacts
required: true
runs:
using: composite
steps:
- uses: arduino/setup-protoc@v1
- name: Install rust toolchain
uses: dtolnay/rust-toolchain@master
with:
toolchain: ${{ inputs.rust-toolchain }}
targets: ${{ inputs.arch }}
components: llvm-tools-preview
- name: Rust Cache
uses: Swatinem/rust-cache@v2
- name: Install Python
uses: actions/setup-python@v4
with:
python-version: '3.10'
- name: Install PyArrow Package
shell: pwsh
run: pip install pyarrow
- name: Install WSL distribution
uses: Vampire/setup-wsl@v2
with:
distribution: Ubuntu-22.04
- name: Install latest nextest release # For integration tests.
if: ${{ inputs.disable-run-tests == 'false' }}
uses: taiki-e/install-action@nextest
- name: Run integration tests
if: ${{ inputs.disable-run-tests == 'false' }}
shell: pwsh
run: make test sqlness-test
- name: Upload sqlness logs
if: ${{ failure() }} # Only upload logs when the integration tests failed.
uses: actions/upload-artifact@v3
with:
name: sqlness-logs
path: ${{ runner.temp }}/greptime-*.log
retention-days: 3
- name: Build greptime binary
shell: pwsh
run: cargo build --profile ${{ inputs.cargo-profile }} --features ${{ inputs.features }} --target ${{ inputs.arch }}
- name: Upload artifacts
uses: ./.github/actions/upload-artifacts
with:
artifacts-dir: ${{ inputs.artifacts-dir }}
target-file: target/${{ inputs.arch }}/${{ inputs.cargo-profile }}/greptime
version: ${{ inputs.version }}

View File

@@ -1,5 +1,5 @@
name: Release artifacts
description: Release artifacts
name: Publish GitHub release
description: Publish GitHub release
inputs:
version:
description: Version to release

View File

@@ -0,0 +1,138 @@
name: Release CN artifacts
description: Release artifacts to CN region
inputs:
src-image-registry:
description: The source image registry to store the images
required: true
default: docker.io
src-image-namespace:
description: The namespace of the source image registry to store the images
required: true
default: greptime
src-image-name:
description: The name of the source image
required: false
default: greptimedb
dst-image-registry:
description: The destination image registry to store the images
required: true
dst-image-namespace:
description: The namespace of the destination image registry to store the images
required: true
default: greptime
dst-image-registry-username:
description: The username to login to the image registry
required: true
dst-image-registry-password:
description: The password to login to the image registry
required: true
version:
description: Version of the artifact
required: true
dev-mode:
description: Enable dev mode, only push standard greptime
required: false
default: 'false'
push-latest-tag:
description: Whether to push the latest tag of the image
required: false
default: 'true'
aws-cn-s3-bucket:
description: S3 bucket to store released artifacts in CN region
required: true
aws-cn-access-key-id:
description: AWS access key id in CN region
required: true
aws-cn-secret-access-key:
description: AWS secret access key in CN region
required: true
aws-cn-region:
description: AWS region in CN
required: true
upload-to-s3:
description: Upload to S3
required: false
default: 'true'
artifacts-dir:
description: Directory to store artifacts
required: false
default: 'artifacts'
update-version-info:
description: Update the version info in S3
required: false
default: 'true'
upload-max-retry-times:
description: Max retry times for uploading artifacts to S3
required: false
default: "20"
upload-retry-timeout:
description: Timeout for uploading artifacts to S3
required: false
default: "30" # minutes
runs:
using: composite
steps:
- name: Download artifacts
uses: actions/download-artifact@v3
with:
path: ${{ inputs.artifacts-dir }}
- name: Release artifacts to cn region
uses: nick-invision/retry@v2
if: ${{ inputs.upload-to-s3 == 'true' }}
env:
AWS_ACCESS_KEY_ID: ${{ inputs.aws-cn-access-key-id }}
AWS_SECRET_ACCESS_KEY: ${{ inputs.aws-cn-secret-access-key }}
AWS_DEFAULT_REGION: ${{ inputs.aws-cn-region }}
UPDATE_VERSION_INFO: ${{ inputs.update-version-info }}
with:
max_attempts: ${{ inputs.upload-max-retry-times }}
timeout_minutes: ${{ inputs.upload-retry-timeout }}
command: |
./.github/scripts/upload-artifacts-to-s3.sh \
${{ inputs.artifacts-dir }} \
${{ inputs.version }} \
${{ inputs.aws-cn-s3-bucket }}
- name: Push greptimedb image from Dockerhub to ACR
shell: bash
env:
DST_REGISTRY_USERNAME: ${{ inputs.dst-image-registry-username }}
DST_REGISTRY_PASSWORD: ${{ inputs.dst-image-registry-password }}
run: |
./.github/scripts/copy-image.sh \
${{ inputs.src-image-registry }}/${{ inputs.src-image-namespace }}/${{ inputs.src-image-name }}:${{ inputs.version }} \
${{ inputs.dst-image-registry }}/${{ inputs.dst-image-namespace }}
- name: Push latest greptimedb image from Dockerhub to ACR
shell: bash
if: ${{ inputs.push-latest-tag == 'true' }}
env:
DST_REGISTRY_USERNAME: ${{ inputs.dst-image-registry-username }}
DST_REGISTRY_PASSWORD: ${{ inputs.dst-image-registry-password }}
run: |
./.github/scripts/copy-image.sh \
${{ inputs.src-image-registry }}/${{ inputs.src-image-namespace }}/${{ inputs.src-image-name }}:latest \
${{ inputs.dst-image-registry }}/${{ inputs.dst-image-namespace }}
- name: Push greptimedb-centos image from DockerHub to ACR
shell: bash
if: ${{ inputs.dev-mode == 'false' }}
env:
DST_REGISTRY_USERNAME: ${{ inputs.dst-image-registry-username }}
DST_REGISTRY_PASSWORD: ${{ inputs.dst-image-registry-password }}
run: |
./.github/scripts/copy-image.sh \
${{ inputs.src-image-registry }}/${{ inputs.src-image-namespace }}/${{ inputs.src-image-name }}-centos:latest \
${{ inputs.dst-image-registry }}/${{ inputs.dst-image-namespace }}
- name: Push greptimedb-centos image from DockerHub to ACR
shell: bash
if: ${{ inputs.dev-mode == 'false' && inputs.push-latest-tag == 'true' }}
env:
DST_REGISTRY_USERNAME: ${{ inputs.dst-image-registry-username }}
DST_REGISTRY_PASSWORD: ${{ inputs.dst-image-registry-password }}
run: |
./.github/scripts/copy-image.sh \
${{ inputs.src-image-registry }}/${{ inputs.src-image-namespace }}/${{ inputs.src-image-name }}-centos:latest \
${{ inputs.dst-image-registry }}/${{ inputs.dst-image-namespace }}

View File

@@ -10,22 +10,6 @@ inputs:
version:
description: Version of the artifact
required: true
release-to-s3-bucket:
description: S3 bucket to store released artifacts
required: true
aws-access-key-id:
description: AWS access key id
required: true
aws-secret-access-key:
description: AWS secret access key
required: true
aws-region:
description: AWS region
required: true
upload-to-s3:
description: Upload to S3
required: false
default: 'true'
working-dir:
description: Working directory to upload the artifacts
required: false
@@ -49,9 +33,21 @@ runs:
working-directory: ${{ inputs.working-dir }}
shell: bash
run: |
tar -zcvf ${{ inputs.artifacts-dir }}.tar.gz ${{ inputs.artifacts-dir }} && \
tar -zcvf ${{ inputs.artifacts-dir }}.tar.gz ${{ inputs.artifacts-dir }}
- name: Calculate checksum
if: runner.os != 'Windows'
working-directory: ${{ inputs.working-dir }}
shell: bash
run: |
echo $(shasum -a 256 ${{ inputs.artifacts-dir }}.tar.gz | cut -f1 -d' ') > ${{ inputs.artifacts-dir }}.sha256sum
- name: Calculate checksum on Windows
if: runner.os == 'Windows'
working-directory: ${{ inputs.working-dir }}
shell: pwsh
run: Get-FileHash ${{ inputs.artifacts-dir }}.tar.gz -Algorithm SHA256 | select -ExpandProperty Hash > ${{ inputs.artifacts-dir }}.sha256sum
# Note: The artifacts will be double zip compressed(related issue: https://github.com/actions/upload-artifact/issues/39).
# However, when we use 'actions/download-artifact@v3' to download the artifacts, it will be automatically unzipped.
- name: Upload artifacts
@@ -65,34 +61,3 @@ runs:
with:
name: ${{ inputs.artifacts-dir }}.sha256sum
path: ${{ inputs.working-dir }}/${{ inputs.artifacts-dir }}.sha256sum
- name: Configure AWS credentials
if: ${{ inputs.upload-to-s3 == 'true' }}
uses: aws-actions/configure-aws-credentials@v2
with:
aws-access-key-id: ${{ inputs.aws-access-key-id }}
aws-secret-access-key: ${{ inputs.aws-secret-access-key }}
aws-region: ${{ inputs.aws-region }}
- name: Upload artifacts to S3
if: ${{ inputs.upload-to-s3 == 'true' }}
uses: nick-invision/retry@v2
with:
max_attempts: 20
timeout_minutes: 5
# The bucket layout will be:
# releases/greptimedb
# ├── v0.1.0
# │ ├── greptime-darwin-amd64-pyo3-v0.1.0.sha256sum
# │ └── greptime-darwin-amd64-pyo3-v0.1.0.tar.gz
# └── v0.2.0
# ├── greptime-darwin-amd64-pyo3-v0.2.0.sha256sum
# └── greptime-darwin-amd64-pyo3-v0.2.0.tar.gz
command: |
cd ${{ inputs.working-dir }} && \
aws s3 cp \
${{ inputs.artifacts-dir }}.tar.gz \
s3://${{ inputs.release-to-s3-bucket }}/releases/greptimedb/${{ inputs.version }}/${{ inputs.artifacts-dir }}.tar.gz && \
aws s3 cp \
${{ inputs.artifacts-dir }}.sha256sum \
s3://${{ inputs.release-to-s3-bucket }}/releases/greptimedb/${{ inputs.version }}/${{ inputs.artifacts-dir }}.sha256sum

47
.github/scripts/copy-image.sh vendored Executable file
View File

@@ -0,0 +1,47 @@
#!/usr/bin/env bash
set -e
set -o pipefail
SRC_IMAGE=$1
DST_REGISTRY=$2
SKOPEO_STABLE_IMAGE="quay.io/skopeo/stable:latest"
# Check if necessary variables are set.
function check_vars() {
for var in DST_REGISTRY_USERNAME DST_REGISTRY_PASSWORD DST_REGISTRY SRC_IMAGE; do
if [ -z "${!var}" ]; then
echo "$var is not set or empty."
echo "Usage: DST_REGISTRY_USERNAME=<your-dst-registry-username> DST_REGISTRY_PASSWORD=<your-dst-registry-password> $0 <dst-registry> <src-image>"
exit 1
fi
done
}
# Copies images from DockerHub to the destination registry.
function copy_images_from_dockerhub() {
# Check if docker is installed.
if ! command -v docker &> /dev/null; then
echo "docker is not installed. Please install docker to continue."
exit 1
fi
# Extract the name and tag of the source image.
IMAGE_NAME=$(echo "$SRC_IMAGE" | sed "s/.*\///")
echo "Copying $SRC_IMAGE to $DST_REGISTRY/$IMAGE_NAME"
docker run "$SKOPEO_STABLE_IMAGE" copy -a docker://"$SRC_IMAGE" \
--dest-creds "$DST_REGISTRY_USERNAME":"$DST_REGISTRY_PASSWORD" \
docker://"$DST_REGISTRY/$IMAGE_NAME"
}
function main() {
check_vars
copy_images_from_dockerhub
}
# Usage example:
# DST_REGISTRY_USERNAME=123 DST_REGISTRY_PASSWORD=456 \
# ./copy-image.sh greptime/greptimedb:v0.4.0 greptime-registry.cn-hangzhou.cr.aliyuncs.com
main

102
.github/scripts/upload-artifacts-to-s3.sh vendored Executable file
View File

@@ -0,0 +1,102 @@
#!/usr/bin/env bash
set -e
set -o pipefail
ARTIFACTS_DIR=$1
VERSION=$2
AWS_S3_BUCKET=$3
RELEASE_DIRS="releases/greptimedb"
GREPTIMEDB_REPO="GreptimeTeam/greptimedb"
# Check if necessary variables are set.
function check_vars() {
for var in AWS_S3_BUCKET VERSION ARTIFACTS_DIR; do
if [ -z "${!var}" ]; then
echo "$var is not set or empty."
echo "Usage: $0 <artifacts-dir> <version> <aws-s3-bucket>"
exit 1
fi
done
}
# Uploads artifacts to AWS S3 bucket.
function upload_artifacts() {
# The bucket layout will be:
# releases/greptimedb
# ├── latest-version.txt
# ├── latest-nightly-version.txt
# ├── v0.1.0
# │ ├── greptime-darwin-amd64-pyo3-v0.1.0.sha256sum
# │ └── greptime-darwin-amd64-pyo3-v0.1.0.tar.gz
# └── v0.2.0
# ├── greptime-darwin-amd64-pyo3-v0.2.0.sha256sum
# └── greptime-darwin-amd64-pyo3-v0.2.0.tar.gz
find "$ARTIFACTS_DIR" -type f \( -name "*.tar.gz" -o -name "*.sha256sum" \) | while IFS= read -r file; do
aws s3 cp \
"$file" "s3://$AWS_S3_BUCKET/$RELEASE_DIRS/$VERSION/$(basename "$file")"
done
}
# Updates the latest version information in AWS S3 if UPDATE_VERSION_INFO is true.
function update_version_info() {
if [ "$UPDATE_VERSION_INFO" == "true" ]; then
# If it's the officail release(like v1.0.0, v1.0.1, v1.0.2, etc.), update latest-version.txt.
if [[ "$VERSION" =~ ^v[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
echo "Updating latest-version.txt"
echo "$VERSION" > latest-version.txt
aws s3 cp \
latest-version.txt "s3://$AWS_S3_BUCKET/$RELEASE_DIRS/latest-version.txt"
fi
# If it's the nightly release, update latest-nightly-version.txt.
if [[ "$VERSION" == *"nightly"* ]]; then
echo "Updating latest-nightly-version.txt"
echo "$VERSION" > latest-nightly-version.txt
aws s3 cp \
latest-nightly-version.txt "s3://$AWS_S3_BUCKET/$RELEASE_DIRS/latest-nightly-version.txt"
fi
fi
}
# Downloads artifacts from Github if DOWNLOAD_ARTIFACTS_FROM_GITHUB is true.
function download_artifacts_from_github() {
if [ "$DOWNLOAD_ARTIFACTS_FROM_GITHUB" == "true" ]; then
# Check if jq is installed.
if ! command -v jq &> /dev/null; then
echo "jq is not installed. Please install jq to continue."
exit 1
fi
# Get the latest release API response.
RELEASES_API_RESPONSE=$(curl -s -H "Accept: application/vnd.github.v3+json" "https://api.github.com/repos/$GREPTIMEDB_REPO/releases/latest")
# Extract download URLs for the artifacts.
# Exclude source code archives which are typically named as 'greptimedb-<version>.zip' or 'greptimedb-<version>.tar.gz'.
ASSET_URLS=$(echo "$RELEASES_API_RESPONSE" | jq -r '.assets[] | select(.name | test("greptimedb-.*\\.(zip|tar\\.gz)$") | not) | .browser_download_url')
# Download each asset.
while IFS= read -r url; do
if [ -n "$url" ]; then
curl -LJO "$url"
echo "Downloaded: $url"
fi
done <<< "$ASSET_URLS"
fi
}
function main() {
check_vars
download_artifacts_from_github
upload_artifacts
update_version_info
}
# Usage example:
# AWS_ACCESS_KEY_ID=<your_access_key_id> \
# AWS_SECRET_ACCESS_KEY=<your_secret_access_key> \
# AWS_DEFAULT_REGION=<your_region> \
# UPDATE_VERSION_INFO=true \
# DOWNLOAD_ARTIFACTS_FROM_GITHUB=false \
# ./upload-artifacts-to-s3.sh <artifacts-dir> <version> <aws-s3-bucket>
main

View File

@@ -17,7 +17,7 @@ env:
jobs:
apidoc:
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- uses: actions/checkout@v3
- uses: arduino/setup-protoc@v1

View File

@@ -16,11 +16,11 @@ on:
description: The runner uses to build linux-amd64 artifacts
default: ec2-c6i.4xlarge-amd64
options:
- ubuntu-latest
- ubuntu-latest-8-cores
- ubuntu-latest-16-cores
- ubuntu-latest-32-cores
- ubuntu-latest-64-cores
- ubuntu-20.04
- ubuntu-20.04-8-cores
- ubuntu-20.04-16-cores
- ubuntu-20.04-32-cores
- ubuntu-20.04-64-cores
- ec2-c6i.xlarge-amd64 # 4C8G
- ec2-c6i.2xlarge-amd64 # 8C16G
- ec2-c6i.4xlarge-amd64 # 16C32G
@@ -78,7 +78,7 @@ jobs:
allocate-runners:
name: Allocate runners
if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
outputs:
linux-amd64-runner: ${{ steps.start-linux-amd64-runner.outputs.label }}
linux-arm64-runner: ${{ steps.start-linux-arm64-runner.outputs.label }}
@@ -164,12 +164,7 @@ jobs:
cargo-profile: ${{ env.CARGO_PROFILE }}
version: ${{ needs.allocate-runners.outputs.version }}
disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
release-to-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
dev-mode: true # Only build the standard greptime binary.
upload-to-s3: false # No need to upload to S3.
working-dir: ${{ env.CHECKOUT_GREPTIMEDB_PATH }}
build-linux-arm64-artifacts:
@@ -198,12 +193,7 @@ jobs:
cargo-profile: ${{ env.CARGO_PROFILE }}
version: ${{ needs.allocate-runners.outputs.version }}
disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
release-to-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
dev-mode: true # Only build the standard greptime binary.
upload-to-s3: false # No need to upload to S3.
working-dir: ${{ env.CHECKOUT_GREPTIMEDB_PATH }}
release-images-to-dockerhub:
@@ -214,7 +204,7 @@ jobs:
build-linux-amd64-artifacts,
build-linux-arm64-artifacts,
]
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
outputs:
build-result: ${{ steps.set-build-result.outputs.build-result }}
steps:
@@ -239,41 +229,44 @@ jobs:
run: |
echo "build-result=success" >> $GITHUB_OUTPUT
release-images-to-acr:
name: Build and push images to ACR
release-cn-artifacts:
name: Release artifacts to CN region
if: ${{ inputs.release_images || github.event_name == 'schedule' }}
needs: [
allocate-runners,
build-linux-amd64-artifacts,
build-linux-arm64-artifacts,
release-images-to-dockerhub,
]
runs-on: ubuntu-latest
# When we push to ACR, it's easy to fail due to some unknown network issues.
# However, we don't want to fail the whole workflow because of this.
# The ACR have daily sync with DockerHub, so don't worry about the image not being updated.
runs-on: ubuntu-20.04
continue-on-error: true
steps:
- uses: actions/checkout@v3
with:
fetch-depth: 0
- name: Build and push images to ACR
uses: ./.github/actions/build-images
- name: Release artifacts to CN region
uses: ./.github/actions/release-cn-artifacts
with:
image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
image-namespace: ${{ vars.IMAGE_NAMESPACE }}
image-name: ${{ env.IMAGE_NAME }}
image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
src-image-registry: docker.io
src-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
src-image-name: greptimedb
dst-image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
dst-image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
dst-image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
dst-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
version: ${{ needs.allocate-runners.outputs.version }}
push-latest-tag: false # Don't push the latest tag to registry.
dev-mode: true # Only build the standard images.
aws-cn-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-cn-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-cn-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-cn-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
dev-mode: true # Only build the standard images(exclude centos images).
push-latest-tag: false # Don't push the latest tag to registry.
update-version-info: false # Don't update the version info in S3.
stop-linux-amd64-runner: # It's always run as the last job in the workflow to make sure that the runner is released.
name: Stop linux-amd64 runner
# Only run this job when the runner is allocated.
if: ${{ always() }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
needs: [
allocate-runners,
build-linux-amd64-artifacts,
@@ -298,7 +291,7 @@ jobs:
name: Stop linux-arm64 runner
# Only run this job when the runner is allocated.
if: ${{ always() }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
needs: [
allocate-runners,
build-linux-arm64-artifacts,
@@ -325,7 +318,7 @@ jobs:
needs: [
release-images-to-dockerhub
]
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL_DEVELOP_CHANNEL }}
steps:
@@ -334,11 +327,11 @@ jobs:
if: ${{ needs.release-images-to-dockerhub.outputs.build-result == 'success' }}
with:
payload: |
{"text": "GreptimeDB ${{ env.NEXT_RELEASE_VERSION }} build successful"}
{"text": "GreptimeDB's ${{ env.NEXT_RELEASE_VERSION }} build has completed successfully."}
- name: Notifiy nightly build failed result
uses: slackapi/slack-github-action@v1.23.0
if: ${{ needs.release-images-to-dockerhub.outputs.build-result != 'success' }}
with:
payload: |
{"text": "GreptimeDB ${{ env.NEXT_RELEASE_VERSION }} build failed, please check 'https://github.com/GreptimeTeam/greptimedb/actions/workflows/${{ env.NEXT_RELEASE_VERSION }}-build.yml'"}
{"text": "GreptimeDB's ${{ env.NEXT_RELEASE_VERSION }} build has failed, please check 'https://github.com/GreptimeTeam/greptimedb/actions/workflows/${{ env.NEXT_RELEASE_VERSION }}-build.yml'."}

View File

@@ -24,13 +24,17 @@ on:
name: CI
concurrency:
group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
cancel-in-progress: true
env:
RUST_TOOLCHAIN: nightly-2023-08-07
jobs:
typos:
name: Spell Check with Typos
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- uses: actions/checkout@v3
- uses: crate-ci/typos@v1.13.10
@@ -38,7 +42,7 @@ jobs:
check:
name: Check
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
timeout-minutes: 60
steps:
- uses: actions/checkout@v3
@@ -56,7 +60,7 @@ jobs:
toml:
name: Toml Check
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
timeout-minutes: 60
steps:
- uses: actions/checkout@v3
@@ -76,7 +80,7 @@ jobs:
runs-on: ${{ matrix.os }}
strategy:
matrix:
os: [ ubuntu-latest-8-cores, windows-latest-8-cores ]
os: [ ubuntu-20.04-8-cores ]
timeout-minutes: 60
steps:
- uses: actions/checkout@v3
@@ -101,7 +105,7 @@ jobs:
fmt:
name: Rustfmt
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
timeout-minutes: 60
steps:
- uses: actions/checkout@v3
@@ -120,7 +124,7 @@ jobs:
clippy:
name: Clippy
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
timeout-minutes: 60
steps:
- uses: actions/checkout@v3
@@ -138,7 +142,7 @@ jobs:
coverage:
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest-8-cores
runs-on: ubuntu-20.04-8-cores
timeout-minutes: 60
steps:
- uses: actions/checkout@v3
@@ -184,43 +188,3 @@ jobs:
flags: rust
fail_ci_if_error: false
verbose: true
test-on-windows:
if: github.event.pull_request.draft == false
runs-on: windows-latest-8-cores
timeout-minutes: 60
steps:
- run: git config --global core.autocrlf false
- uses: actions/checkout@v3
- uses: arduino/setup-protoc@v1
with:
repo-token: ${{ secrets.GITHUB_TOKEN }}
- name: Install Rust toolchain
uses: dtolnay/rust-toolchain@master
with:
toolchain: ${{ env.RUST_TOOLCHAIN }}
components: llvm-tools-preview
- name: Rust Cache
uses: Swatinem/rust-cache@v2
- name: Install Cargo Nextest
uses: taiki-e/install-action@nextest
- name: Install Python
uses: actions/setup-python@v4
with:
python-version: '3.10'
- name: Install PyArrow Package
run: pip install pyarrow
- name: Install WSL distribution
uses: Vampire/setup-wsl@v2
with:
distribution: Ubuntu-22.04
- name: Running tests
run: cargo nextest run -F pyo3_backend,dashboard
env:
RUST_BACKTRACE: 1
CARGO_INCREMENTAL: 0
GT_S3_BUCKET: ${{ secrets.S3_BUCKET }}
GT_S3_ACCESS_KEY_ID: ${{ secrets.S3_ACCESS_KEY_ID }}
GT_S3_ACCESS_KEY: ${{ secrets.S3_ACCESS_KEY }}
GT_S3_REGION: ${{ secrets.S3_REGION }}
UNITTEST_LOG_DIR: "__unittest_logs"

View File

@@ -11,7 +11,7 @@ on:
jobs:
doc_issue:
if: github.event.label.name == 'doc update required'
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- name: create an issue in doc repo
uses: dacbd/create-issue-action@main
@@ -25,7 +25,7 @@ jobs:
${{ github.event.issue.html_url || github.event.pull_request.html_url }}
cloud_issue:
if: github.event.label.name == 'cloud followup required'
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- name: create an issue in cloud repo
uses: dacbd/create-issue-action@main

View File

@@ -30,7 +30,7 @@ name: CI
jobs:
typos:
name: Spell Check with Typos
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- uses: actions/checkout@v3
- uses: crate-ci/typos@v1.13.10
@@ -38,33 +38,33 @@ jobs:
check:
name: Check
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- run: 'echo "No action required"'
fmt:
name: Rustfmt
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- run: 'echo "No action required"'
clippy:
name: Clippy
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- run: 'echo "No action required"'
coverage:
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- run: 'echo "No action required"'
sqlness:
name: Sqlness Test
if: github.event.pull_request.draft == false
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- run: 'echo "No action required"'

View File

@@ -8,7 +8,7 @@ on:
types: [opened, synchronize, reopened, ready_for_review]
jobs:
license-header-check:
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
name: license-header-check
steps:
- uses: actions/checkout@v2

View File

@@ -14,11 +14,11 @@ on:
description: The runner uses to build linux-amd64 artifacts
default: ec2-c6i.2xlarge-amd64
options:
- ubuntu-latest
- ubuntu-latest-8-cores
- ubuntu-latest-16-cores
- ubuntu-latest-32-cores
- ubuntu-latest-64-cores
- ubuntu-20.04
- ubuntu-20.04-8-cores
- ubuntu-20.04-16-cores
- ubuntu-20.04-32-cores
- ubuntu-20.04-64-cores
- ec2-c6i.xlarge-amd64 # 4C8G
- ec2-c6i.2xlarge-amd64 # 8C16G
- ec2-c6i.4xlarge-amd64 # 16C32G
@@ -70,7 +70,7 @@ jobs:
allocate-runners:
name: Allocate runners
if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
outputs:
linux-amd64-runner: ${{ steps.start-linux-amd64-runner.outputs.label }}
linux-arm64-runner: ${{ steps.start-linux-arm64-runner.outputs.label }}
@@ -147,10 +147,6 @@ jobs:
cargo-profile: ${{ env.CARGO_PROFILE }}
version: ${{ needs.allocate-runners.outputs.version }}
disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
release-to-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
build-linux-arm64-artifacts:
name: Build linux-arm64 artifacts
@@ -170,10 +166,6 @@ jobs:
cargo-profile: ${{ env.CARGO_PROFILE }}
version: ${{ needs.allocate-runners.outputs.version }}
disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
release-to-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
release-images-to-dockerhub:
name: Build and push images to DockerHub
@@ -183,7 +175,7 @@ jobs:
build-linux-amd64-artifacts,
build-linux-arm64-artifacts,
]
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
outputs:
nightly-build-result: ${{ steps.set-nightly-build-result.outputs.nightly-build-result }}
steps:
@@ -206,15 +198,14 @@ jobs:
run: |
echo "nightly-build-result=success" >> $GITHUB_OUTPUT
release-images-to-acr:
name: Build and push images to ACR
release-cn-artifacts:
name: Release artifacts to CN region
if: ${{ inputs.release_images || github.event_name == 'schedule' }}
needs: [
allocate-runners,
build-linux-amd64-artifacts,
build-linux-arm64-artifacts,
release-images-to-dockerhub,
]
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
# When we push to ACR, it's easy to fail due to some unknown network issues.
# However, we don't want to fail the whole workflow because of this.
# The ACR have daily sync with DockerHub, so don't worry about the image not being updated.
@@ -224,21 +215,30 @@ jobs:
with:
fetch-depth: 0
- name: Build and push images to ACR
uses: ./.github/actions/build-images
- name: Release artifacts to CN region
uses: ./.github/actions/release-cn-artifacts
with:
image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
image-namespace: ${{ vars.IMAGE_NAMESPACE }}
image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
src-image-registry: docker.io
src-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
src-image-name: greptimedb
dst-image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
dst-image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
dst-image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
dst-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
version: ${{ needs.allocate-runners.outputs.version }}
push-latest-tag: false # Don't push the latest tag to registry.
aws-cn-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-cn-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-cn-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-cn-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
dev-mode: false
update-version-info: false # Don't update version info in S3.
push-latest-tag: false # Don't push the latest tag to registry.
stop-linux-amd64-runner: # It's always run as the last job in the workflow to make sure that the runner is released.
name: Stop linux-amd64 runner
# Only run this job when the runner is allocated.
if: ${{ always() }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
needs: [
allocate-runners,
build-linux-amd64-artifacts,
@@ -263,7 +263,7 @@ jobs:
name: Stop linux-arm64 runner
# Only run this job when the runner is allocated.
if: ${{ always() }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
needs: [
allocate-runners,
build-linux-arm64-artifacts,
@@ -290,7 +290,7 @@ jobs:
needs: [
release-images-to-dockerhub
]
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
env:
SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL_DEVELOP_CHANNEL }}
steps:
@@ -299,11 +299,11 @@ jobs:
if: ${{ needs.release-images-to-dockerhub.outputs.nightly-build-result == 'success' }}
with:
payload: |
{"text": "GreptimeDB nightly build successful"}
{"text": "GreptimeDB's ${{ env.NEXT_RELEASE_VERSION }} build has completed successfully."}
- name: Notifiy nightly build failed result
uses: slackapi/slack-github-action@v1.23.0
if: ${{ needs.release-images-to-dockerhub.outputs.nightly-build-result != 'success' }}
with:
payload: |
{"text": "GreptimeDB nightly build failed, please check 'https://github.com/GreptimeTeam/greptimedb/actions/workflows/nightly-build.yml'"}
{"text": "GreptimeDB's ${{ env.NEXT_RELEASE_VERSION }} build has failed, please check 'https://github.com/GreptimeTeam/greptimedb/actions/workflows/${{ env.NEXT_RELEASE_VERSION }}-build.yml'."}

82
.github/workflows/nightly-ci.yml vendored Normal file
View File

@@ -0,0 +1,82 @@
# Nightly CI: runs tests every night for our second tier plaforms (Windows)
on:
schedule:
- cron: '0 23 * * 1-5'
workflow_dispatch:
name: Nightly CI
concurrency:
group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
cancel-in-progress: true
env:
RUST_TOOLCHAIN: nightly-2023-08-07
jobs:
sqlness:
name: Sqlness Test
runs-on: ${{ matrix.os }}
strategy:
matrix:
os: [ windows-latest-8-cores ]
timeout-minutes: 60
steps:
- uses: actions/checkout@v4.1.0
- uses: arduino/setup-protoc@v1
with:
repo-token: ${{ secrets.GITHUB_TOKEN }}
- uses: dtolnay/rust-toolchain@master
with:
toolchain: ${{ env.RUST_TOOLCHAIN }}
- name: Rust Cache
uses: Swatinem/rust-cache@v2
- name: Run sqlness
run: cargo sqlness
- name: Upload sqlness logs
if: always()
uses: actions/upload-artifact@v3
with:
name: sqlness-logs
path: ${{ runner.temp }}/greptime-*.log
retention-days: 3
test-on-windows:
runs-on: windows-latest-8-cores
timeout-minutes: 60
steps:
- run: git config --global core.autocrlf false
- uses: actions/checkout@v4.1.0
- uses: arduino/setup-protoc@v1
with:
repo-token: ${{ secrets.GITHUB_TOKEN }}
- name: Install Rust toolchain
uses: dtolnay/rust-toolchain@master
with:
toolchain: ${{ env.RUST_TOOLCHAIN }}
components: llvm-tools-preview
- name: Rust Cache
uses: Swatinem/rust-cache@v2
- name: Install Cargo Nextest
uses: taiki-e/install-action@nextest
- name: Install Python
uses: actions/setup-python@v4
with:
python-version: '3.10'
- name: Install PyArrow Package
run: pip install pyarrow
- name: Install WSL distribution
uses: Vampire/setup-wsl@v2
with:
distribution: Ubuntu-22.04
- name: Running tests
run: cargo nextest run -F pyo3_backend,dashboard
env:
RUST_BACKTRACE: 1
CARGO_INCREMENTAL: 0
GT_S3_BUCKET: ${{ secrets.S3_BUCKET }}
GT_S3_ACCESS_KEY_ID: ${{ secrets.S3_ACCESS_KEY_ID }}
GT_S3_ACCESS_KEY: ${{ secrets.S3_ACCESS_KEY }}
GT_S3_REGION: ${{ secrets.S3_REGION }}
UNITTEST_LOG_DIR: "__unittest_logs"

View File

@@ -10,7 +10,7 @@ on:
jobs:
check:
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
timeout-minutes: 10
steps:
- uses: thehanimo/pr-title-checker@v1.3.4
@@ -19,7 +19,7 @@ jobs:
pass_on_octokit_error: false
configuration_path: ".github/pr-title-checker-config.json"
breaking:
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
timeout-minutes: 10
steps:
- uses: thehanimo/pr-title-checker@v1.3.4

View File

@@ -0,0 +1,85 @@
name: Release dev-builder images
on:
workflow_dispatch: # Allows you to run this workflow manually.
inputs:
version:
description: Version of the dev-builder
required: false
default: latest
release_dev_builder_ubuntu_image:
type: boolean
description: Release dev-builder-ubuntu image
required: false
default: false
release_dev_builder_centos_image:
type: boolean
description: Release dev-builder-centos image
required: false
default: false
release_dev_builder_android_image:
type: boolean
description: Release dev-builder-android image
required: false
default: false
jobs:
release-dev-builder-images:
name: Release dev builder images
if: ${{ inputs.release_dev_builder_ubuntu_image || inputs.release_dev_builder_centos_image || inputs.release_dev_builder_android_image }} # Only manually trigger this job.
runs-on: ubuntu-20.04-16-cores
steps:
- name: Checkout
uses: actions/checkout@v3
with:
fetch-depth: 0
- name: Build and push dev builder images
uses: ./.github/actions/build-dev-builder-images
with:
version: ${{ inputs.version }}
dockerhub-image-registry-username: ${{ secrets.DOCKERHUB_USERNAME }}
dockerhub-image-registry-token: ${{ secrets.DOCKERHUB_TOKEN }}
build-dev-builder-ubuntu: ${{ inputs.release_dev_builder_ubuntu_image }}
build-dev-builder-centos: ${{ inputs.release_dev_builder_centos_image }}
build-dev-builder-android: ${{ inputs.release_dev_builder_android_image }}
release-dev-builder-images-cn: # Note: Be careful issue: https://github.com/containers/skopeo/issues/1874 and we decide to use the latest stable skopeo container.
name: Release dev builder images to CN region
runs-on: ubuntu-20.04
needs: [
release-dev-builder-images
]
steps:
- name: Push dev-builder-ubuntu image
shell: bash
if: ${{ inputs.release_dev_builder_ubuntu_image }}
env:
DST_REGISTRY_USERNAME: ${{ secrets.ALICLOUD_USERNAME }}
DST_REGISTRY_PASSWORD: ${{ secrets.ALICLOUD_PASSWORD }}
run: |
docker run quay.io/skopeo/stable:latest copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ inputs.version }} \
--dest-creds "$DST_REGISTRY_USERNAME":"$DST_REGISTRY_PASSWORD" \
docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-ubuntu:${{ inputs.version }}
- name: Push dev-builder-centos image
shell: bash
if: ${{ inputs.release_dev_builder_centos_image }}
env:
DST_REGISTRY_USERNAME: ${{ secrets.ALICLOUD_USERNAME }}
DST_REGISTRY_PASSWORD: ${{ secrets.ALICLOUD_PASSWORD }}
run: |
docker run quay.io/skopeo/stable:latest copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:${{ inputs.version }} \
--dest-creds "$DST_REGISTRY_USERNAME":"$DST_REGISTRY_PASSWORD" \
docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-centos:${{ inputs.version }}
- name: Push dev-builder-android image
shell: bash
if: ${{ inputs.release_dev_builder_android_image }}
env:
DST_REGISTRY_USERNAME: ${{ secrets.ALICLOUD_USERNAME }}
DST_REGISTRY_PASSWORD: ${{ secrets.ALICLOUD_PASSWORD }}
run: |
docker run quay.io/skopeo/stable:latest copy -a docker://docker.io/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:${{ inputs.version }} \
--dest-creds "$DST_REGISTRY_USERNAME":"$DST_REGISTRY_PASSWORD" \
docker://${{ vars.ACR_IMAGE_REGISTRY }}/${{ vars.IMAGE_NAMESPACE }}/dev-builder-android:${{ inputs.version }}

View File

@@ -18,11 +18,11 @@ on:
description: The runner uses to build linux-amd64 artifacts
default: ec2-c6i.4xlarge-amd64
options:
- ubuntu-latest
- ubuntu-latest-8-cores
- ubuntu-latest-16-cores
- ubuntu-latest-32-cores
- ubuntu-latest-64-cores
- ubuntu-20.04
- ubuntu-20.04-8-cores
- ubuntu-20.04-16-cores
- ubuntu-20.04-32-cores
- ubuntu-20.04-64-cores
- ec2-c6i.xlarge-amd64 # 4C8G
- ec2-c6i.2xlarge-amd64 # 8C16G
- ec2-c6i.4xlarge-amd64 # 16C32G
@@ -63,7 +63,12 @@ on:
description: Build macos artifacts
required: false
default: false
release_artifacts:
build_windows_artifacts:
type: boolean
description: Build Windows artifacts
required: false
default: false
publish_github_release:
type: boolean
description: Create GitHub release and upload artifacts
required: false
@@ -73,11 +78,6 @@ on:
description: Build and push images to DockerHub and ACR
required: false
default: false
release_dev_builder_image:
type: boolean
description: Release dev-builder image
required: false
default: false
# Use env variables to control all the release process.
env:
@@ -91,17 +91,18 @@ env:
# The scheduled version is '${{ env.NEXT_RELEASE_VERSION }}-nightly-YYYYMMDD', like v0.2.0-nigthly-20230313;
NIGHTLY_RELEASE_PREFIX: nightly
# Note: The NEXT_RELEASE_VERSION should be modified manually by every formal release.
NEXT_RELEASE_VERSION: v0.4.0
NEXT_RELEASE_VERSION: v0.5.0
jobs:
allocate-runners:
name: Allocate runners
if: ${{ github.repository == 'GreptimeTeam/greptimedb' }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
outputs:
linux-amd64-runner: ${{ steps.start-linux-amd64-runner.outputs.label }}
linux-arm64-runner: ${{ steps.start-linux-arm64-runner.outputs.label }}
macos-runner: ${{ inputs.macos_runner || vars.DEFAULT_MACOS_RUNNER }}
windows-runner: windows-latest-8-cores
# The following EC2 resource id will be used for resource releasing.
linux-amd64-ec2-runner-label: ${{ steps.start-linux-amd64-runner.outputs.label }}
@@ -177,10 +178,6 @@ jobs:
cargo-profile: ${{ env.CARGO_PROFILE }}
version: ${{ needs.allocate-runners.outputs.version }}
disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
release-to-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
build-linux-arm64-artifacts:
name: Build linux-arm64 artifacts
@@ -200,10 +197,6 @@ jobs:
cargo-profile: ${{ env.CARGO_PROFILE }}
version: ${{ needs.allocate-runners.outputs.version }}
disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
release-to-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
build-macos-artifacts:
name: Build macOS artifacts
@@ -245,11 +238,43 @@ jobs:
features: ${{ matrix.features }}
version: ${{ needs.allocate-runners.outputs.version }}
disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
release-to-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
artifacts-dir: ${{ matrix.artifacts-dir-prefix }}-${{ needs.allocate-runners.outputs.version }}
aws-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
build-windows-artifacts:
name: Build Windows artifacts
strategy:
fail-fast: false
matrix:
include:
- os: ${{ needs.allocate-runners.outputs.windows-runner }}
arch: x86_64-pc-windows-msvc
features: servers/dashboard
artifacts-dir-prefix: greptime-windows-amd64
- os: ${{ needs.allocate-runners.outputs.windows-runner }}
arch: x86_64-pc-windows-msvc
features: pyo3_backend,servers/dashboard
artifacts-dir-prefix: greptime-windows-amd64-pyo3
runs-on: ${{ matrix.os }}
needs: [
allocate-runners,
]
if: ${{ inputs.build_windows_artifacts || github.event_name == 'push' || github.event_name == 'schedule' }}
steps:
- run: git config --global core.autocrlf false
- uses: actions/checkout@v3
with:
fetch-depth: 0
- uses: ./.github/actions/build-windows-artifacts
with:
arch: ${{ matrix.arch }}
rust-toolchain: ${{ env.RUST_TOOLCHAIN }}
cargo-profile: ${{ env.CARGO_PROFILE }}
features: ${{ matrix.features }}
version: ${{ needs.allocate-runners.outputs.version }}
disable-run-tests: ${{ env.DISABLE_RUN_TESTS }}
artifacts-dir: ${{ matrix.artifacts-dir-prefix }}-${{ needs.allocate-runners.outputs.version }}
release-images-to-dockerhub:
name: Build and push images to DockerHub
@@ -274,15 +299,14 @@ jobs:
image-registry-password: ${{ secrets.DOCKERHUB_TOKEN }}
version: ${{ needs.allocate-runners.outputs.version }}
release-images-to-acr:
name: Build and push images to ACR
release-cn-artifacts:
name: Release artifacts to CN region
if: ${{ inputs.release_images || github.event_name == 'push' || github.event_name == 'schedule' }}
needs: [
allocate-runners,
build-linux-amd64-artifacts,
build-linux-arm64-artifacts,
release-images-to-dockerhub,
]
runs-on: ubuntu-2004-16-cores
runs-on: ubuntu-20.04
# When we push to ACR, it's easy to fail due to some unknown network issues.
# However, we don't want to fail the whole workflow because of this.
# The ACR have daily sync with DockerHub, so don't worry about the image not being updated.
@@ -292,18 +316,28 @@ jobs:
with:
fetch-depth: 0
- name: Build and push images to ACR
uses: ./.github/actions/build-images
- name: Release artifacts to CN region
uses: ./.github/actions/release-cn-artifacts
with:
image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
image-namespace: ${{ vars.IMAGE_NAMESPACE }}
image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
src-image-registry: docker.io
src-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
src-image-name: greptimedb
dst-image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
dst-image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
dst-image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
dst-image-namespace: ${{ vars.IMAGE_NAMESPACE }}
version: ${{ needs.allocate-runners.outputs.version }}
aws-cn-s3-bucket: ${{ vars.AWS_RELEASE_BUCKET }}
aws-cn-access-key-id: ${{ secrets.AWS_CN_ACCESS_KEY_ID }}
aws-cn-secret-access-key: ${{ secrets.AWS_CN_SECRET_ACCESS_KEY }}
aws-cn-region: ${{ vars.AWS_RELEASE_BUCKET_REGION }}
dev-mode: false
update-version-info: true
push-latest-tag: true
release-artifacts:
publish-github-release:
name: Create GitHub release and upload artifacts
if: ${{ inputs.release_artifacts || github.event_name == 'push' || github.event_name == 'schedule' }}
if: ${{ inputs.publish_github_release || github.event_name == 'push' || github.event_name == 'schedule' }}
needs: [
allocate-runners,
build-linux-amd64-artifacts,
@@ -311,36 +345,17 @@ jobs:
build-macos-artifacts,
release-images-to-dockerhub,
]
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
steps:
- uses: actions/checkout@v3
with:
fetch-depth: 0
- name: Release artifacts
uses: ./.github/actions/release-artifacts
- name: Publish GitHub release
uses: ./.github/actions/publish-github-release
with:
version: ${{ needs.allocate-runners.outputs.version }}
release-dev-builder-image:
name: Release dev builder image
if: ${{ inputs.release_dev_builder_image }} # Only manually trigger this job.
runs-on: ubuntu-latest-16-cores
steps:
- name: Checkout
uses: actions/checkout@v3
with:
fetch-depth: 0
- name: Build and push dev builder image
uses: ./.github/actions/build-dev-builder-image
with:
dockerhub-image-registry-username: ${{ secrets.DOCKERHUB_USERNAME }}
dockerhub-image-registry-token: ${{ secrets.DOCKERHUB_TOKEN }}
acr-image-registry: ${{ vars.ACR_IMAGE_REGISTRY }}
acr-image-registry-username: ${{ secrets.ALICLOUD_USERNAME }}
acr-image-registry-password: ${{ secrets.ALICLOUD_PASSWORD }}
### Stop runners ###
# It's very necessary to split the job of releasing runners into 'stop-linux-amd64-runner' and 'stop-linux-arm64-runner'.
# Because we can terminate the specified EC2 instance immediately after the job is finished without uncessary waiting.
@@ -348,7 +363,7 @@ jobs:
name: Stop linux-amd64 runner
# Only run this job when the runner is allocated.
if: ${{ always() }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
needs: [
allocate-runners,
build-linux-amd64-artifacts,
@@ -373,7 +388,7 @@ jobs:
name: Stop linux-arm64 runner
# Only run this job when the runner is allocated.
if: ${{ always() }}
runs-on: ubuntu-latest
runs-on: ubuntu-20.04
needs: [
allocate-runners,
build-linux-arm64-artifacts,

1597
Cargo.lock generated

File diff suppressed because it is too large Load Diff

View File

@@ -2,15 +2,17 @@
members = [
"benchmarks",
"src/api",
"src/auth",
"src/catalog",
"src/client",
"src/cmd",
"src/common/base",
"src/common/catalog",
"src/common/config",
"src/common/datasource",
"src/common/error",
"src/common/function",
"src/common/function-macro",
"src/common/macro",
"src/common/greptimedb-telemetry",
"src/common/grpc",
"src/common/grpc-expr",
@@ -28,15 +30,16 @@ members = [
"src/common/version",
"src/datanode",
"src/datatypes",
"src/file-table-engine",
"src/file-engine",
"src/frontend",
"src/log-store",
"src/meta-client",
"src/meta-srv",
"src/mito",
"src/mito2",
"src/object-store",
"src/operator",
"src/partition",
"src/plugins",
"src/promql",
"src/query",
"src/script",
@@ -46,90 +49,101 @@ members = [
"src/storage",
"src/store-api",
"src/table",
"src/table-procedure",
"tests-integration",
"tests/runner",
]
resolver = "2"
[workspace.package]
version = "0.3.2"
version = "0.4.0"
edition = "2021"
license = "Apache-2.0"
[workspace.dependencies]
aquamarine = "0.3"
arrow = { version = "43.0" }
etcd-client = "0.11"
arrow-array = "43.0"
arrow-flight = "43.0"
arrow-schema = { version = "43.0", features = ["serde"] }
async-stream = "0.3"
async-trait = "0.1"
chrono = { version = "0.4", features = ["serde"] }
datafusion = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "2ceb7f927c40787773fdc466d6a4b79f3a6c0001" }
datafusion-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "2ceb7f927c40787773fdc466d6a4b79f3a6c0001" }
datafusion-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "2ceb7f927c40787773fdc466d6a4b79f3a6c0001" }
datafusion-optimizer = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "2ceb7f927c40787773fdc466d6a4b79f3a6c0001" }
datafusion-physical-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "2ceb7f927c40787773fdc466d6a4b79f3a6c0001" }
datafusion-sql = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "2ceb7f927c40787773fdc466d6a4b79f3a6c0001" }
datafusion-substrait = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "2ceb7f927c40787773fdc466d6a4b79f3a6c0001" }
datafusion = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "b6f3b28b6fe91924cc8dd3d83726b766f2a706ec" }
datafusion-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "b6f3b28b6fe91924cc8dd3d83726b766f2a706ec" }
datafusion-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "b6f3b28b6fe91924cc8dd3d83726b766f2a706ec" }
datafusion-optimizer = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "b6f3b28b6fe91924cc8dd3d83726b766f2a706ec" }
datafusion-physical-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "b6f3b28b6fe91924cc8dd3d83726b766f2a706ec" }
datafusion-sql = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "b6f3b28b6fe91924cc8dd3d83726b766f2a706ec" }
datafusion-substrait = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "b6f3b28b6fe91924cc8dd3d83726b766f2a706ec" }
derive_builder = "0.12"
etcd-client = "0.11"
futures = "0.3"
futures-util = "0.3"
greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "940694cfd05c1e93c1dd7aab486184c9e2853098" }
greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "1f1dd532a111e3834cc3019c5605e2993ffb9dc3" }
humantime-serde = "1.1"
itertools = "0.10"
lazy_static = "1.4"
meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "abbd357c1e193cd270ea65ee7652334a150b628f" }
metrics = "0.20"
moka = "0.12"
once_cell = "1.18"
opentelemetry-proto = { version = "0.2", features = ["gen-tonic", "metrics"] }
parquet = "43.0"
paste = "1.0"
prost = "0.11"
raft-engine = { git = "https://github.com/tikv/raft-engine.git", rev = "22dfb426cd994602b57725ef080287d3e53db479" }
rand = "0.8"
regex = "1.8"
reqwest = { version = "0.11", default-features = false, features = [
"json",
"rustls-tls-native-roots",
"stream",
] }
serde = { version = "1.0", features = ["derive"] }
serde_json = "1.0"
smallvec = "1"
snafu = { version = "0.7", features = ["backtraces"] }
sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "c3814f08afa19786b13d72b1731a1e8b3cac4ab9", features = [
sqlparser = { git = "https://github.com/GreptimeTeam/sqlparser-rs.git", rev = "296a4f6c73b129d6f565a42a2e5e53c6bc2b9da4", features = [
"visitor",
] }
strum = { version = "0.25", features = ["derive"] }
tempfile = "3"
tokio = { version = "1.28", features = ["full"] }
tokio-util = { version = "0.7", features = ["io-util", "compat"] }
toml = "0.7"
tonic = { version = "0.9", features = ["tls"] }
uuid = { version = "1", features = ["serde", "v4", "fast-rng"] }
metrics = "0.20"
meter-core = { git = "https://github.com/GreptimeTeam/greptime-meter.git", rev = "abbd357c1e193cd270ea65ee7652334a150b628f" }
## workspaces members
api = { path = "src/api" }
auth = { path = "src/auth" }
catalog = { path = "src/catalog" }
client = { path = "src/client" }
cmd = { path = "src/cmd" }
common-base = { path = "src/common/base" }
common-catalog = { path = "src/common/catalog" }
common-config = { path = "src/common/config" }
common-datasource = { path = "src/common/datasource" }
common-error = { path = "src/common/error" }
common-function = { path = "src/common/function" }
common-function-macro = { path = "src/common/function-macro" }
common-greptimedb-telemetry = { path = "src/common/greptimedb-telemetry" }
common-grpc = { path = "src/common/grpc" }
common-grpc-expr = { path = "src/common/grpc-expr" }
common-macro = { path = "src/common/macro" }
common-mem-prof = { path = "src/common/mem-prof" }
common-meta = { path = "src/common/meta" }
common-pprof = { path = "src/common/pprof" }
common-procedure = { path = "src/common/procedure" }
common-procedure-test = { path = "src/common/procedure-test" }
common-pprof = { path = "src/common/pprof" }
common-query = { path = "src/common/query" }
common-recordbatch = { path = "src/common/recordbatch" }
common-runtime = { path = "src/common/runtime" }
substrait = { path = "src/common/substrait" }
common-telemetry = { path = "src/common/telemetry" }
common-test-util = { path = "src/common/test-util" }
common-time = { path = "src/common/time" }
common-version = { path = "src/common/version" }
datanode = { path = "src/datanode" }
datatypes = { path = "src/datatypes" }
file-table-engine = { path = "src/file-table-engine" }
file-engine = { path = "src/file-engine" }
frontend = { path = "src/frontend" }
log-store = { path = "src/log-store" }
meta-client = { path = "src/meta-client" }
@@ -137,7 +151,9 @@ meta-srv = { path = "src/meta-srv" }
mito = { path = "src/mito" }
mito2 = { path = "src/mito2" }
object-store = { path = "src/object-store" }
operator = { path = "src/operator" }
partition = { path = "src/partition" }
plugins = { path = "src/plugins" }
promql = { path = "src/promql" }
query = { path = "src/query" }
script = { path = "src/script" }
@@ -146,8 +162,8 @@ session = { path = "src/session" }
sql = { path = "src/sql" }
storage = { path = "src/storage" }
store-api = { path = "src/store-api" }
substrait = { path = "src/common/substrait" }
table = { path = "src/table" }
table-procedure = { path = "src/table-procedure" }
[workspace.dependencies.meter-macros]
git = "https://github.com/GreptimeTeam/greptime-meter.git"

View File

@@ -12,6 +12,8 @@ BUILDX_BUILDER_NAME ?= gtbuilder
BASE_IMAGE ?= ubuntu
RUST_TOOLCHAIN ?= $(shell cat rust-toolchain.toml | grep channel | cut -d'"' -f2)
CARGO_REGISTRY_CACHE ?= ${HOME}/.cargo/registry
ARCH := $(shell uname -m | sed 's/x86_64/amd64/' | sed 's/aarch64/arm64/')
OUTPUT_DIR := $(shell if [ "$(RELEASE)" = "true" ]; then echo "release"; elif [ ! -z "$(CARGO_PROFILE)" ]; then echo "$(CARGO_PROFILE)" ; else echo "debug"; fi)
# The arguments for running integration tests.
ETCD_VERSION ?= v3.5.9
@@ -43,35 +45,60 @@ ifneq ($(strip $(TARGET)),)
CARGO_BUILD_OPTS += --target ${TARGET}
endif
ifneq ($(strip $(RELEASE)),)
CARGO_BUILD_OPTS += --release
endif
ifeq ($(BUILDX_MULTI_PLATFORM_BUILD), true)
BUILDX_MULTI_PLATFORM_BUILD_OPTS := --platform linux/amd64,linux/arm64 --push
else
BUILDX_MULTI_PLATFORM_BUILD_OPTS := -o type=docker
endif
ifneq ($(strip $(CARGO_BUILD_EXTRA_OPTS)),)
CARGO_BUILD_OPTS += ${CARGO_BUILD_EXTRA_OPTS}
endif
##@ Build
.PHONY: build
build: ## Build debug version greptime. If USE_DEV_BUILDER is true, the binary will be built in dev-builder.
ifeq ($(USE_DEV_BUILDER), true)
docker run --network=host \
-v ${PWD}:/greptimedb -v ${CARGO_REGISTRY_CACHE}:/root/.cargo/registry \
-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder:latest \
make build CARGO_PROFILE=${CARGO_PROFILE} FEATURES=${FEATURES} TARGET_DIR=${TARGET_DIR}
else
cargo build ${CARGO_BUILD_OPTS}
endif
build: ## Build debug version greptime.
cargo ${CARGO_EXTENSION} build ${CARGO_BUILD_OPTS}
.PHONY: release
release: ## Build release version greptime. If USE_DEV_BUILDER is true, the binary will be built in dev-builder.
ifeq ($(USE_DEV_BUILDER), true)
.POHNY: build-by-dev-builder
build-by-dev-builder: ## Build greptime by dev-builder.
docker run --network=host \
-v ${PWD}:/greptimedb -v ${CARGO_REGISTRY_CACHE}:/root/.cargo/registry \
-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder:latest \
make release CARGO_PROFILE=${CARGO_PROFILE} FEATURES=${FEATURES} TARGET_DIR=${TARGET_DIR}
else
cargo build --release ${CARGO_BUILD_OPTS}
endif
-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:latest \
make build \
CARGO_EXTENSION="${CARGO_EXTENSION}" \
CARGO_PROFILE=${CARGO_PROFILE} \
FEATURES=${FEATURES} \
TARGET_DIR=${TARGET_DIR} \
TARGET=${TARGET} \
RELEASE=${RELEASE} \
CARGO_BUILD_EXTRA_OPTS="${CARGO_BUILD_EXTRA_OPTS}"
.PHONY: build-android-bin
build-android-bin: ## Build greptime binary for android.
docker run --network=host \
-v ${PWD}:/greptimedb -v ${CARGO_REGISTRY_CACHE}:/root/.cargo/registry \
-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-android:latest \
make build \
CARGO_EXTENSION="ndk --platform 23 -t aarch64-linux-android" \
CARGO_PROFILE=release \
FEATURES="${FEATURES}" \
TARGET_DIR="${TARGET_DIR}" \
TARGET="${TARGET}" \
RELEASE="${RELEASE}" \
CARGO_BUILD_EXTRA_OPTS="--bin greptime --no-default-features"
.PHONY: strip-android-bin
strip-android-bin: build-android-bin ## Strip greptime binary for android.
docker run --network=host \
-v ${PWD}:/greptimedb \
-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-android:latest \
bash -c '$${NDK_ROOT}/toolchains/llvm/prebuilt/linux-x86_64/bin/llvm-strip /greptimedb/target/aarch64-linux-android/release/greptime'
.PHONY: clean
clean: ## Clean the project.
@@ -90,30 +117,27 @@ check-toml: ## Check all TOML files.
taplo format --check
.PHONY: docker-image
docker-image: multi-platform-buildx ## Build docker image.
docker-image: build-by-dev-builder ## Build docker image.
mkdir -p ${ARCH} && \
cp ./target/${OUTPUT_DIR}/greptime ${ARCH}/greptime && \
docker build -f docker/ci/${BASE_IMAGE}/Dockerfile -t ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/greptimedb:${IMAGE_TAG} . && \
rm -r ${ARCH}
.PHONY: docker-image-buildx
docker-image-buildx: multi-platform-buildx ## Build docker image by buildx.
docker buildx build --builder ${BUILDX_BUILDER_NAME} \
--build-arg="CARGO_PROFILE=${CARGO_PROFILE}" --build-arg="FEATURES=${FEATURES}" \
-f docker/${BASE_IMAGE}/Dockerfile \
--build-arg="CARGO_PROFILE=${CARGO_PROFILE}" \
--build-arg="FEATURES=${FEATURES}" \
--build-arg="OUTPUT_DIR=${OUTPUT_DIR}" \
-f docker/buildx/${BASE_IMAGE}/Dockerfile \
-t ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/greptimedb:${IMAGE_TAG} ${BUILDX_MULTI_PLATFORM_BUILD_OPTS} .
.PHONY: build-greptime-by-buildx
build-greptime-by-buildx: multi-platform-buildx ## Build greptime binary by docker buildx. The binary will be copied to the current directory.
docker buildx build --builder ${BUILDX_BUILDER_NAME} \
--target=builder \
--build-arg="CARGO_PROFILE=${CARGO_PROFILE}" --build-arg="FEATURES=${FEATURES}" \
-f docker/${BASE_IMAGE}/Dockerfile \
-t ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/greptimedb-builder:${IMAGE_TAG} ${BUILDX_MULTI_PLATFORM_BUILD_OPTS} .
docker run --rm -v ${PWD}:/data \
--entrypoint cp ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/greptimedb-builder:${IMAGE_TAG} \
/out/target/${CARGO_PROFILE}/greptime /data/greptime
.PHONY: dev-builder
dev-builder: multi-platform-buildx ## Build dev-builder image.
docker buildx build --builder ${BUILDX_BUILDER_NAME} \
--build-arg="RUST_TOOLCHAIN=${RUST_TOOLCHAIN}" \
-f docker/dev-builder/Dockerfile \
-t ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder:${IMAGE_TAG} ${BUILDX_MULTI_PLATFORM_BUILD_OPTS} .
-f docker/dev-builder/${BASE_IMAGE}/Dockerfile \
-t ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:${IMAGE_TAG} ${BUILDX_MULTI_PLATFORM_BUILD_OPTS} .
.PHONY: multi-platform-buildx
multi-platform-buildx: ## Create buildx multi-platform builder.
@@ -155,7 +179,7 @@ stop-etcd: ## Stop single node etcd for testing purpose.
run-it-in-container: start-etcd ## Run integration tests in dev-builder.
docker run --network=host \
-v ${PWD}:/greptimedb -v ${CARGO_REGISTRY_CACHE}:/root/.cargo/registry -v /tmp:/tmp \
-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder:latest \
-w /greptimedb ${IMAGE_REGISTRY}/${IMAGE_NAMESPACE}/dev-builder-${BASE_IMAGE}:latest \
make test sqlness-test BUILD_JOBS=${BUILD_JOBS}
##@ General

View File

@@ -27,6 +27,14 @@
<a href="https://greptime.com/slack"><img src="https://img.shields.io/badge/slack-GreptimeDB-0abd59?logo=slack" alt="slack" /></a>
</p>
## Upcoming Event
Come and meet us in **KubeCon + CloudNativeCon North America 2023!**
<p align="center">
<picture>
<img alt="KubeCon + CloudNativeCon North Logo" src="./docs/banner/KCCNC_NA_2023_1000x200_Email Banner.png" width="800px">
</picture>
</p>
## What is GreptimeDB
GreptimeDB is an open-source time-series database with a special focus on
@@ -96,11 +104,11 @@ Or if you built from docker:
docker run -p 4002:4002 -v "$(pwd):/tmp/greptimedb" greptime/greptimedb standalone start
```
Please see [the online document site](https://docs.greptime.com/getting-started/overview#install-greptimedb) for more installation options and [operations info](https://docs.greptime.com/user-guide/operations/overview).
Please see the online document site for more installation options and [operations info](https://docs.greptime.com/user-guide/operations/overview).
### Get started
Read the [complete getting started guide](https://docs.greptime.com/getting-started/overview#connect) on our [official document site](https://docs.greptime.com/).
Read the [complete getting started guide](https://docs.greptime.com/getting-started/try-out-greptimedb) on our [official document site](https://docs.greptime.com/).
To write and query data, GreptimeDB is compatible with multiple [protocols and clients](https://docs.greptime.com/user-guide/clients/overview).

View File

@@ -6,8 +6,10 @@ license.workspace = true
[dependencies]
arrow.workspace = true
chrono.workspace = true
clap = { version = "4.0", features = ["derive"] }
client = { workspace = true }
futures-util.workspace = true
indicatif = "0.17.1"
itertools.workspace = true
parquet.workspace = true

View File

@@ -27,16 +27,16 @@ use arrow::record_batch::RecordBatch;
use clap::Parser;
use client::api::v1::column::Values;
use client::api::v1::{
Column, ColumnDataType, ColumnDef, CreateTableExpr, InsertRequest, InsertRequests,
Column, ColumnDataType, ColumnDef, CreateTableExpr, InsertRequest, InsertRequests, SemanticType,
};
use client::{Client, Database, DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
use client::{Client, Database, Output, DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
use futures_util::TryStreamExt;
use indicatif::{MultiProgress, ProgressBar, ProgressStyle};
use parquet::arrow::arrow_reader::ParquetRecordBatchReaderBuilder;
use tokio::task::JoinSet;
const CATALOG_NAME: &str = "greptime";
const SCHEMA_NAME: &str = "public";
const TABLE_NAME: &str = "nyc_taxi";
#[derive(Parser)]
#[command(name = "NYC benchmark runner")]
@@ -74,7 +74,12 @@ fn get_file_list<P: AsRef<Path>>(path: P) -> Vec<PathBuf> {
.collect()
}
fn new_table_name() -> String {
format!("nyc_taxi_{}", chrono::Utc::now().timestamp())
}
async fn write_data(
table_name: &str,
batch_size: usize,
db: &Database,
path: PathBuf,
@@ -104,8 +109,7 @@ async fn write_data(
}
let (columns, row_count) = convert_record_batch(record_batch);
let request = InsertRequest {
table_name: TABLE_NAME.to_string(),
region_number: 0,
table_name: table_name.to_string(),
columns,
row_count,
};
@@ -114,7 +118,7 @@ async fn write_data(
};
let now = Instant::now();
let _ = db.insert(requests).await.unwrap();
db.insert(requests).await.unwrap();
let elapsed = now.elapsed();
total_rpc_elapsed_ms += elapsed.as_millis();
progress_bar.inc(row_count as _);
@@ -132,6 +136,11 @@ fn convert_record_batch(record_batch: RecordBatch) -> (Vec<Column>, u32) {
for (array, field) in record_batch.columns().iter().zip(fields.iter()) {
let (values, datatype) = build_values(array);
let semantic_type = match field.name().as_str() {
"VendorID" => SemanticType::Tag,
"tpep_pickup_datetime" => SemanticType::Timestamp,
_ => SemanticType::Field,
};
let column = Column {
column_name: field.name().clone(),
@@ -142,8 +151,7 @@ fn convert_record_batch(record_batch: RecordBatch) -> (Vec<Column>, u32) {
.map(|bitmap| bitmap.buffer().as_slice().to_vec())
.unwrap_or_default(),
datatype: datatype.into(),
// datatype and semantic_type are set to default
..Default::default()
semantic_type: semantic_type as i32,
};
columns.push(column);
}
@@ -189,7 +197,7 @@ fn build_values(column: &ArrayRef) -> (Values, ColumnDataType) {
let values = array.values();
(
Values {
ts_microsecond_values: values.to_vec(),
timestamp_microsecond_values: values.to_vec(),
..Default::default()
},
ColumnDataType::TimestampMicrosecond,
@@ -244,156 +252,193 @@ fn is_record_batch_full(batch: &RecordBatch) -> bool {
batch.columns().iter().all(|col| col.null_count() == 0)
}
fn create_table_expr() -> CreateTableExpr {
fn create_table_expr(table_name: &str) -> CreateTableExpr {
CreateTableExpr {
catalog_name: CATALOG_NAME.to_string(),
schema_name: SCHEMA_NAME.to_string(),
table_name: TABLE_NAME.to_string(),
table_name: table_name.to_string(),
desc: "".to_string(),
column_defs: vec![
ColumnDef {
name: "VendorID".to_string(),
datatype: ColumnDataType::Int64 as i32,
data_type: ColumnDataType::Int64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Tag as i32,
comment: String::new(),
},
ColumnDef {
name: "tpep_pickup_datetime".to_string(),
datatype: ColumnDataType::TimestampMicrosecond as i32,
is_nullable: true,
data_type: ColumnDataType::TimestampMicrosecond as i32,
is_nullable: false,
default_constraint: vec![],
semantic_type: SemanticType::Timestamp as i32,
comment: String::new(),
},
ColumnDef {
name: "tpep_dropoff_datetime".to_string(),
datatype: ColumnDataType::TimestampMicrosecond as i32,
data_type: ColumnDataType::TimestampMicrosecond as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "passenger_count".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "trip_distance".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "RatecodeID".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "store_and_fwd_flag".to_string(),
datatype: ColumnDataType::String as i32,
data_type: ColumnDataType::String as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "PULocationID".to_string(),
datatype: ColumnDataType::Int64 as i32,
data_type: ColumnDataType::Int64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "DOLocationID".to_string(),
datatype: ColumnDataType::Int64 as i32,
data_type: ColumnDataType::Int64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "payment_type".to_string(),
datatype: ColumnDataType::Int64 as i32,
data_type: ColumnDataType::Int64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "fare_amount".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "extra".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "mta_tax".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "tip_amount".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "tolls_amount".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "improvement_surcharge".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "total_amount".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "congestion_surcharge".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
ColumnDef {
name: "airport_fee".to_string(),
datatype: ColumnDataType::Float64 as i32,
data_type: ColumnDataType::Float64 as i32,
is_nullable: true,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
],
time_index: "tpep_pickup_datetime".to_string(),
primary_keys: vec!["VendorID".to_string()],
create_if_not_exists: false,
create_if_not_exists: true,
table_options: Default::default(),
region_numbers: vec![0],
table_id: None,
engine: "mito".to_string(),
}
}
fn query_set() -> HashMap<String, String> {
fn query_set(table_name: &str) -> HashMap<String, String> {
HashMap::from([
(
"count_all".to_string(),
format!("SELECT COUNT(*) FROM {TABLE_NAME};"),
format!("SELECT COUNT(*) FROM {table_name};"),
),
(
"fare_amt_by_passenger".to_string(),
format!("SELECT passenger_count, MIN(fare_amount), MAX(fare_amount), SUM(fare_amount) FROM {TABLE_NAME} GROUP BY passenger_count"),
format!("SELECT passenger_count, MIN(fare_amount), MAX(fare_amount), SUM(fare_amount) FROM {table_name} GROUP BY passenger_count"),
)
])
}
async fn do_write(args: &Args, db: &Database) {
async fn do_write(args: &Args, db: &Database, table_name: &str) {
let mut file_list = get_file_list(args.path.clone().expect("Specify data path in argument"));
let mut write_jobs = JoinSet::new();
let create_table_result = db.create(create_table_expr()).await;
let create_table_result = db.create(create_table_expr(table_name)).await;
println!("Create table result: {create_table_result:?}");
let progress_bar_style = ProgressStyle::with_template(
@@ -411,8 +456,10 @@ async fn do_write(args: &Args, db: &Database) {
let db = db.clone();
let mpb = multi_progress_bar.clone();
let pb_style = progress_bar_style.clone();
let _ = write_jobs
.spawn(async move { write_data(batch_size, &db, path, mpb, pb_style).await });
let table_name = table_name.to_string();
let _ = write_jobs.spawn(async move {
write_data(&table_name, batch_size, &db, path, mpb, pb_style).await
});
}
}
while write_jobs.join_next().await.is_some() {
@@ -421,24 +468,32 @@ async fn do_write(args: &Args, db: &Database) {
let db = db.clone();
let mpb = multi_progress_bar.clone();
let pb_style = progress_bar_style.clone();
let _ = write_jobs
.spawn(async move { write_data(batch_size, &db, path, mpb, pb_style).await });
let table_name = table_name.to_string();
let _ = write_jobs.spawn(async move {
write_data(&table_name, batch_size, &db, path, mpb, pb_style).await
});
}
}
}
async fn do_query(num_iter: usize, db: &Database) {
for (query_name, query) in query_set() {
async fn do_query(num_iter: usize, db: &Database, table_name: &str) {
for (query_name, query) in query_set(table_name) {
println!("Running query: {query}");
for i in 0..num_iter {
let now = Instant::now();
let _res = db.sql(&query).await.unwrap();
let res = db.sql(&query).await.unwrap();
match res {
Output::AffectedRows(_) | Output::RecordBatches(_) => (),
Output::Stream(stream) => {
stream.try_collect::<Vec<_>>().await.unwrap();
}
}
let elapsed = now.elapsed();
println!(
"query {}, iteration {}: {}ms",
query_name,
i,
elapsed.as_millis()
elapsed.as_millis(),
);
}
}
@@ -455,13 +510,14 @@ fn main() {
.block_on(async {
let client = Client::with_urls(vec![&args.endpoint]);
let db = Database::new(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, client);
let table_name = new_table_name();
if !args.skip_write {
do_write(&args, &db).await;
do_write(&args, &db, &table_name).await;
}
if !args.skip_read {
do_query(args.iter_num, &db).await;
do_query(args.iter_num, &db, &table_name).await;
}
})
}

View File

@@ -1,7 +1,5 @@
# Node running mode, see `standalone.example.toml`.
mode = "distributed"
# Whether to use in-memory catalog, see `standalone.example.toml`.
enable_memory_catalog = false
# The datanode identifier, should be unique.
node_id = 42
# gRPC server address, "127.0.0.1:3001" by default.
@@ -10,19 +8,24 @@ rpc_addr = "127.0.0.1:3001"
rpc_hostname = "127.0.0.1"
# The number of gRPC server worker threads, 8 by default.
rpc_runtime_size = 8
# Start services after regions have obtained leases.
# It will block the datanode start if it can't receive leases in the heartbeat from metasrv.
require_lease_before_startup = false
[heartbeat]
# Interval for sending heartbeat messages to the Metasrv in milliseconds, 5000 by default.
interval_millis = 5000
# Interval for sending heartbeat messages to the Metasrv, 3 seconds by default.
interval = "3s"
# Metasrv client options.
[meta_client_options]
[meta_client]
# Metasrv address list.
metasrv_addrs = ["127.0.0.1:3002"]
# Operation timeout in milliseconds, 3000 by default.
timeout_millis = 3000
# Connect server timeout in milliseconds, 5000 by default.
connect_timeout_millis = 5000
# Heartbeat timeout, 500 milliseconds by default.
heartbeat_timeout = "500ms"
# Operation timeout, 3 seconds by default.
timeout = "3s"
# Connect server timeout, 1 second by default.
connect_timeout = "1s"
# `TCP_NODELAY` option for accepted connections, true by default.
tcp_nodelay = true
@@ -44,6 +47,12 @@ type = "File"
# TTL for all tables. Disabled by default.
# global_ttl = "7d"
# Cache configuration for object storage such as 'S3' etc.
# The local file cache directory
# cache_path = "/path/local_cache"
# The local file cache capacity in bytes.
# cache_capacity = "256MB"
# Compaction options, see `standalone.example.toml`.
[storage.compaction]
max_inflight_tasks = 4
@@ -57,8 +66,6 @@ max_purge_tasks = 32
checkpoint_margin = 10
# Region manifest logs and checkpoints gc execution duration
gc_duration = '10m'
# Whether to try creating a manifest checkpoint on region opening
checkpoint_on_startup = false
# Storage flush options
[storage.flush]
@@ -73,10 +80,27 @@ auto_flush_interval = "1h"
# Global write buffer size for all regions.
global_write_buffer_size = "1GB"
# Procedure storage options, see `standalone.example.toml`.
[procedure]
max_retry_times = 3
retry_delay = "500ms"
# Mito engine options
[[region_engine]]
[region_engine.mito]
# Number of region workers
num_workers = 8
# Request channel size of each worker
worker_channel_size = 128
# Max batch size for a worker to handle requests
worker_request_batch_size = 64
# Number of meta action updated to trigger a new checkpoint for the manifest
manifest_checkpoint_distance = 10
# Manifest compression type
manifest_compress_type = "Uncompressed"
# Max number of running background jobs
max_background_jobs = 4
# Interval to auto flush a region if it has not flushed yet.
auto_flush_interval = "1h"
# Global write buffer size for all regions.
global_write_buffer_size = "1GB"
# Global write buffer size threshold to reject write requests (default 2G).
global_write_buffer_reject_size = "2GB"
# Log options
# [logging]

View File

@@ -2,68 +2,67 @@
mode = "distributed"
[heartbeat]
# Interval for sending heartbeat task to the Metasrv in milliseconds, 5000 by default.
interval_millis = 5000
# Interval for retry sending heartbeat task in milliseconds, 5000 by default.
retry_interval_millis = 5000
# Interval for sending heartbeat task to the Metasrv, 5 seconds by default.
interval = "5s"
# Interval for retry sending heartbeat task, 5 seconds by default.
retry_interval = "5s"
# HTTP server options, see `standalone.example.toml`.
[http_options]
[http]
addr = "127.0.0.1:4000"
timeout = "30s"
body_limit = "64MB"
# gRPC server options, see `standalone.example.toml`.
[grpc_options]
[grpc]
addr = "127.0.0.1:4001"
runtime_size = 8
# MySQL server options, see `standalone.example.toml`.
[mysql_options]
[mysql]
enable = true
addr = "127.0.0.1:4002"
runtime_size = 2
# MySQL server TLS options, see `standalone.example.toml`.
[mysql_options.tls]
[mysql.tls]
mode = "disable"
cert_path = ""
key_path = ""
# PostgresSQL server options, see `standalone.example.toml`.
[postgres_options]
[postgres]
enable = true
addr = "127.0.0.1:4003"
runtime_size = 2
# PostgresSQL server TLS options, see `standalone.example.toml`.
[postgres_options.tls]
[postgres.tls]
mode = "disable"
cert_path = ""
key_path = ""
# OpenTSDB protocol options, see `standalone.example.toml`.
[opentsdb_options]
[opentsdb]
enable = true
addr = "127.0.0.1:4242"
runtime_size = 2
# InfluxDB protocol options, see `standalone.example.toml`.
[influxdb_options]
[influxdb]
enable = true
# Prometheus remote storage options, see `standalone.example.toml`.
[prom_store_options]
[prom_store]
enable = true
# Prometheus protocol options, see `standalone.example.toml`.
[prometheus_options]
addr = "127.0.0.1:4004"
# Metasrv client options, see `datanode.example.toml`.
[meta_client_options]
[meta_client]
metasrv_addrs = ["127.0.0.1:3002"]
timeout_millis = 3000
timeout = "3s"
# DDL timeouts options.
ddl_timeout_millis = 10000
connect_timeout_millis = 5000
ddl_timeout = "10s"
connect_timeout = "1s"
tcp_nodelay = true
# Log options, see `standalone.example.toml`

View File

@@ -6,8 +6,6 @@ bind_addr = "127.0.0.1:3002"
server_addr = "127.0.0.1:3002"
# Etcd server address, "127.0.0.1:2379" by default.
store_addr = "127.0.0.1:2379"
# Datanode lease in seconds, 15 seconds by default.
datanode_lease_secs = 15
# Datanode selector type.
# - "LeaseBased" (default value).
# - "LoadBased"
@@ -26,7 +24,7 @@ enable_telemetry = true
# Procedure storage options.
[procedure]
# Procedure max retry time.
max_retry_times = 3
max_retry_times = 12
# Initial retry delay of procedures, increases exponentially
retry_delay = "500ms"
@@ -34,6 +32,6 @@ retry_delay = "500ms"
# [datanode]
# # Datanode client options.
# [datanode.client_options]
# timeout_millis = 10000
# connect_timeout_millis = 10000
# timeout = "10s"
# connect_timeout = "10s"
# tcp_nodelay = true

View File

@@ -1,12 +1,10 @@
# Node running mode, "standalone" or "distributed".
mode = "standalone"
# Whether to use in-memory catalog, `false` by default.
enable_memory_catalog = false
# Whether to enable greptimedb telemetry, true by default.
enable_telemetry = true
# HTTP server options.
[http_options]
[http]
# Server address, "127.0.0.1:4000" by default.
addr = "127.0.0.1:4000"
# HTTP request timeout, 30s by default.
@@ -16,21 +14,23 @@ timeout = "30s"
body_limit = "64MB"
# gRPC server options.
[grpc_options]
[grpc]
# Server address, "127.0.0.1:4001" by default.
addr = "127.0.0.1:4001"
# The number of server worker threads, 8 by default.
runtime_size = 8
# MySQL server options.
[mysql_options]
[mysql]
# Whether to enable
enable = true
# Server address, "127.0.0.1:4002" by default.
addr = "127.0.0.1:4002"
# The number of server worker threads, 2 by default.
runtime_size = 2
# MySQL server TLS options.
[mysql_options.tls]
[mysql.tls]
# TLS mode, refer to https://www.postgresql.org/docs/current/libpq-ssl.html
# - "disable" (default value)
# - "prefer"
@@ -44,14 +44,16 @@ cert_path = ""
key_path = ""
# PostgresSQL server options.
[postgres_options]
[postgres]
# Whether to enable
enable = true
# Server address, "127.0.0.1:4003" by default.
addr = "127.0.0.1:4003"
# The number of server worker threads, 2 by default.
runtime_size = 2
# PostgresSQL server TLS options, see `[mysql_options.tls]` section.
[postgres_options.tls]
[postgres.tls]
# TLS mode.
mode = "disable"
# certificate file path.
@@ -60,31 +62,26 @@ cert_path = ""
key_path = ""
# OpenTSDB protocol options.
[opentsdb_options]
[opentsdb]
# Whether to enable
enable = true
# OpenTSDB telnet API server address, "127.0.0.1:4242" by default.
addr = "127.0.0.1:4242"
# The number of server worker threads, 2 by default.
runtime_size = 2
# InfluxDB protocol options.
[influxdb_options]
[influxdb]
# Whether to enable InfluxDB protocol in HTTP API, true by default.
enable = true
# Prometheus remote storage options
[prom_store_options]
[prom_store]
# Whether to enable Prometheus remote write and read in HTTP API, true by default.
enable = true
# Prometheus protocol options
[prometheus_options]
# Prometheus API server address, "127.0.0.1:4004" by default.
addr = "127.0.0.1:4004"
# WAL options.
[wal]
# WAL data directory
# dir = "/tmp/greptimedb/wal"
# WAL file size in bytes.
file_size = "256MB"
# WAL purge threshold.
@@ -96,6 +93,20 @@ read_batch_size = 128
# Whether to sync log file after every write.
sync_write = false
# Metadata storage options.
[metadata_store]
# Kv file size in bytes.
file_size = "256MB"
# Kv purge threshold.
purge_threshold = "4GB"
# Procedure storage options.
[procedure]
# Procedure max retry time.
max_retry_times = 3
# Initial retry delay of procedures, increases exponentially
retry_delay = "500ms"
# Storage options.
[storage]
# The working home directory.
@@ -104,6 +115,10 @@ data_home = "/tmp/greptimedb/"
type = "File"
# TTL for all tables. Disabled by default.
# global_ttl = "7d"
# Cache configuration for object storage such as 'S3' etc.
# cache_path = "/path/local_cache"
# The local file cache capacity in bytes.
# cache_capacity = "256MB"
# Compaction options.
[storage.compaction]
@@ -121,8 +136,6 @@ max_purge_tasks = 32
checkpoint_margin = 10
# Region manifest logs and checkpoints gc execution duration
gc_duration = '10m'
# Whether to try creating a manifest checkpoint on region opening
checkpoint_on_startup = false
# Storage flush options
[storage.flush]
@@ -137,13 +150,6 @@ auto_flush_interval = "1h"
# Global write buffer size for all regions.
global_write_buffer_size = "1GB"
# Procedure storage options.
[procedure]
# Procedure max retry time.
max_retry_times = 3
# Initial retry delay of procedures, increases exponentially
retry_delay = "500ms"
# Log options
# [logging]
# Specify logs directory.

View File

@@ -2,6 +2,7 @@ FROM centos:7 as builder
ARG CARGO_PROFILE
ARG FEATURES
ARG OUTPUT_DIR
ENV LANG en_US.utf8
WORKDIR /greptimedb
@@ -13,7 +14,8 @@ RUN yum install -y epel-release \
openssl-devel \
centos-release-scl \
rh-python38 \
rh-python38-python-devel
rh-python38-python-devel \
which
# Install protoc
RUN curl -LO https://github.com/protocolbuffers/protobuf/releases/download/v3.15.8/protoc-3.15.8-linux-x86_64.zip
@@ -35,17 +37,18 @@ RUN --mount=target=.,rw \
# Export the binary to the clean image.
FROM centos:7 as base
ARG CARGO_PROFILE
ARG OUTPUT_DIR
RUN yum install -y epel-release \
openssl \
openssl-devel \
centos-release-scl \
rh-python38 \
rh-python38-python-devel
rh-python38-python-devel \
which
WORKDIR /greptime
COPY --from=builder /out/target/${CARGO_PROFILE}/greptime /greptime/bin/
COPY --from=builder /out/target/${OUTPUT_DIR}/greptime /greptime/bin/
ENV PATH /greptime/bin/:$PATH
ENTRYPOINT ["greptime"]

View File

@@ -1,11 +1,17 @@
FROM ubuntu:22.04 as builder
FROM ubuntu:20.04 as builder
ARG CARGO_PROFILE
ARG FEATURES
ARG OUTPUT_DIR
ENV LANG en_US.utf8
WORKDIR /greptimedb
# Add PPA for Python 3.10.
RUN apt-get update && \
DEBIAN_FRONTEND=noninteractive apt-get install -y software-properties-common && \
add-apt-repository ppa:deadsnakes/ppa -y
# Install dependencies.
RUN --mount=type=cache,target=/var/cache/apt \
apt-get update && apt-get install -y \
@@ -25,7 +31,7 @@ RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- --no-mo
ENV PATH /root/.cargo/bin/:$PATH
# Build the project in release mode.
RUN --mount=target=.,rw \
RUN --mount=target=. \
--mount=type=cache,target=/root/.cargo/registry \
make build \
CARGO_PROFILE=${CARGO_PROFILE} \
@@ -36,7 +42,7 @@ RUN --mount=target=.,rw \
# TODO(zyy17): Maybe should use the more secure container image.
FROM ubuntu:22.04 as base
ARG CARGO_PROFILE
ARG OUTPUT_DIR
RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get \
-y install ca-certificates \
@@ -50,7 +56,7 @@ COPY ./docker/python/requirements.txt /etc/greptime/requirements.txt
RUN python3 -m pip install -r /etc/greptime/requirements.txt
WORKDIR /greptime
COPY --from=builder /out/target/${CARGO_PROFILE}/greptime /greptime/bin/
COPY --from=builder /out/target/${OUTPUT_DIR}/greptime /greptime/bin/
ENV PATH /greptime/bin/:$PATH
ENTRYPOINT ["greptime"]

View File

@@ -0,0 +1,41 @@
FROM --platform=linux/amd64 saschpe/android-ndk:34-jdk17.0.8_7-ndk25.2.9519653-cmake3.22.1
ENV LANG en_US.utf8
WORKDIR /greptimedb
# Rename libunwind to libgcc
RUN cp ${NDK_ROOT}/toolchains/llvm/prebuilt/linux-x86_64/lib64/clang/14.0.7/lib/linux/aarch64/libunwind.a ${NDK_ROOT}/toolchains/llvm/prebuilt/linux-x86_64/lib64/clang/14.0.7/lib/linux/aarch64/libgcc.a
# Install dependencies.
RUN apt-get update && apt-get install -y \
libssl-dev \
protobuf-compiler \
curl \
git \
build-essential \
pkg-config \
python3 \
python3-dev \
python3-pip \
&& pip3 install --upgrade pip \
&& pip3 install pyarrow
# Trust workdir
RUN git config --global --add safe.directory /greptimedb
# Install Rust.
SHELL ["/bin/bash", "-c"]
RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- --no-modify-path --default-toolchain none -y
ENV PATH /root/.cargo/bin/:$PATH
# Add android toolchains
ARG RUST_TOOLCHAIN
RUN rustup toolchain install ${RUST_TOOLCHAIN}
RUN rustup target add aarch64-linux-android
# Install cargo-ndk
RUN cargo install cargo-ndk
ENV ANDROID_NDK_HOME $NDK_ROOT
# Builder entrypoint.
CMD ["cargo", "ndk", "--platform", "23", "-t", "aarch64-linux-android", "build", "--bin", "greptime", "--profile", "release", "--no-default-features"]

View File

@@ -0,0 +1,29 @@
FROM centos:7 as builder
ENV LANG en_US.utf8
# Install dependencies
RUN ulimit -n 1024000 && yum groupinstall -y 'Development Tools'
RUN yum install -y epel-release \
openssl \
openssl-devel \
centos-release-scl \
rh-python38 \
rh-python38-python-devel \
which
# Install protoc
RUN curl -LO https://github.com/protocolbuffers/protobuf/releases/download/v3.15.8/protoc-3.15.8-linux-x86_64.zip
RUN unzip protoc-3.15.8-linux-x86_64.zip -d /usr/local/
# Install Rust
SHELL ["/bin/bash", "-c"]
RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- --no-modify-path --default-toolchain none -y
ENV PATH /opt/rh/rh-python38/root/usr/bin:/usr/local/bin:/root/.cargo/bin/:$PATH
# Install Rust toolchains.
ARG RUST_TOOLCHAIN
RUN rustup toolchain install ${RUST_TOOLCHAIN}
# Install nextest.
RUN cargo install cargo-nextest --locked

View File

@@ -1,8 +1,13 @@
FROM ubuntu:22.04
FROM ubuntu:20.04
ENV LANG en_US.utf8
WORKDIR /greptimedb
# Add PPA for Python 3.10.
RUN apt-get update && \
DEBIAN_FRONTEND=noninteractive apt-get install -y software-properties-common && \
add-apt-repository ppa:deadsnakes/ppa -y
# Install dependencies.
RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y \
libssl-dev \

Binary file not shown.

After

Width:  |  Height:  |  Size: 51 KiB

View File

@@ -0,0 +1,61 @@
# TSBS benchmark - v0.4.0
## Environment
### Local
| | |
| ------ | ---------------------------------- |
| CPU | AMD Ryzen 7 7735HS (8 core 3.2GHz) |
| Memory | 32GB |
| Disk | SOLIDIGM SSDPFKNU010TZ |
| OS | Ubuntu 22.04.2 LTS |
### Aliyun amd64
| | |
| ------- | -------------- |
| Machine | ecs.g7.4xlarge |
| CPU | 16 core |
| Memory | 64GB |
| Disk | 100G |
| OS | Ubuntu 22.04 |
### Aliyun arm64
| | |
| ------- | ----------------- |
| Machine | ecs.g8y.4xlarge |
| CPU | 16 core |
| Memory | 64GB |
| Disk | 100G |
| OS | Ubuntu 22.04 ARM |
## Write performance
| Environment | Ingest raterows/s |
| ------------------ | --------------------- |
| Local | 365280.60 |
| Aliyun g7.4xlarge | 341368.72 |
| Aliyun g8y.4xlarge | 320907.29 |
## Query performance
| Query type | Local (ms) | Aliyun g7.4xlarge (ms) | Aliyun g8y.4xlarge (ms) |
| --------------------- | ---------- | ---------------------- | ----------------------- |
| cpu-max-all-1 | 50.70 | 31.46 | 47.61 |
| cpu-max-all-8 | 262.16 | 129.26 | 152.43 |
| double-groupby-1 | 2512.71 | 1408.19 | 1586.10 |
| double-groupby-5 | 3896.15 | 2304.29 | 2585.29 |
| double-groupby-all | 5404.67 | 3337.61 | 3773.91 |
| groupby-orderby-limit | 3786.98 | 2065.72 | 2312.57 |
| high-cpu-1 | 71.96 | 37.29 | 54.01 |
| high-cpu-all | 9468.75 | 7595.69 | 8467.46 |
| lastpoint | 13379.43 | 11253.76 | 12949.40 |
| single-groupby-1-1-1 | 20.72 | 12.16 | 13.35 |
| single-groupby-1-1-12 | 28.53 | 15.67 | 21.62 |
| single-groupby-1-8-1 | 72.23 | 37.90 | 43.52 |
| single-groupby-5-1-1 | 26.75 | 15.59 | 17.48 |
| single-groupby-5-1-12 | 45.41 | 22.90 | 31.96 |
| single-groupby-5-8-1 | 107.96 | 59.76 | 69.58 |

View File

@@ -0,0 +1,90 @@
---
Feature Name: Update Metadata in single transaction
Tracking Issue: https://github.com/GreptimeTeam/greptimedb/issues/1715
Date: 2023-08-13
Author: "Feng Yangsen <fengys1996@gmail.com>, Xu Wenkang <wenymedia@gmail.com>"
---
# Summary
Update Metadata in single transaction.
# Motivation
Currently, multiple transactions are involved during the procedure. This implementation is inefficient, and it's hard to make data consistent. Therefore, We can update multiple metadata in a single transaction.
# Details
Now we have the following table metadata keys:
**TableInfo**
```rust
// __table_info/{table_id}
pub struct TableInfoKey {
table_id: TableId,
}
pub struct TableInfoValue {
pub table_info: RawTableInfo,
version: u64,
}
```
**TableRoute**
```rust
// __table_route/{table_id}
pub struct NextTableRouteKey {
table_id: TableId,
}
pub struct TableRoute {
pub region_routes: Vec<RegionRoute>,
}
```
**DatanodeTable**
```rust
// __table_route/{datanode_id}/{table_id}
pub struct DatanodeTableKey {
datanode_id: DatanodeId,
table_id: TableId,
}
pub struct DatanodeTableValue {
pub table_id: TableId,
pub regions: Vec<RegionNumber>,
version: u64,
}
```
**TableNameKey**
```rust
// __table_name/{CatalogName}/{SchemaName}/{TableName}
pub struct TableNameKey<'a> {
pub catalog: &'a str,
pub schema: &'a str,
pub table: &'a str,
}
pub struct TableNameValue {
table_id: TableId,
}
```
These table metadata only updates in the following operations.
## Region Failover
It needs to update `TableRoute` key and `DatanodeTable` keys. If the `TableRoute` equals the Snapshot of `TableRoute` submitting the Failover task, then we can safely update these keys.
After submitting Failover tasks to acquire locks for execution, the `TableRoute` may be updated by another task. After acquiring the lock, we can get the latest `TableRoute` again and then execute it if needed.
## Create Table DDL
Creates all of the above keys. `TableRoute`, `TableInfo`, should be empty.
The **TableNameKey**'s lock will be held by the procedure framework.
## Drop Table DDL
`TableInfoKey` and `NextTableRouteKey` will be added with `__removed-` prefix, and the other above keys will be deleted. The transaction will not compare any keys.
## Alter Table DDL
1. Rename table, updates `TableInfo` and `TableName`. Compares `TableInfo`, and the new `TableNameKey` should be empty, and TableInfo should equal the Snapshot when submitting DDL.
The old and new **TableNameKey**'s lock will be held by the procedure framework.
2. Alter table, updates `TableInfo`. `TableInfo` should equal the Snapshot when submitting DDL.

View File

@@ -7,6 +7,7 @@ license.workspace = true
[dependencies]
common-base = { workspace = true }
common-error = { workspace = true }
common-macro = { workspace = true }
common-time = { workspace = true }
datatypes = { workspace = true }
greptime-proto.workspace = true
@@ -16,3 +17,6 @@ tonic.workspace = true
[build-dependencies]
tonic-build = "0.9"
[dev-dependencies]
paste = "1.0"

View File

@@ -16,14 +16,16 @@ use std::any::Any;
use common_error::ext::ErrorExt;
use common_error::status_code::StatusCode;
use common_macro::stack_trace_debug;
use datatypes::prelude::ConcreteDataType;
use snafu::prelude::*;
use snafu::Location;
pub type Result<T> = std::result::Result<T, Error>;
#[derive(Debug, Snafu)]
#[derive(Snafu)]
#[snafu(visibility(pub))]
#[stack_trace_debug]
pub enum Error {
#[snafu(display("Unknown proto column datatype: {}", datatype))]
UnknownColumnDataType { datatype: i32, location: Location },
@@ -34,22 +36,14 @@ pub enum Error {
location: Location,
},
#[snafu(display(
"Failed to convert column default constraint, column: {}, source: {}",
column,
source
))]
#[snafu(display("Failed to convert column default constraint, column: {}", column))]
ConvertColumnDefaultConstraint {
column: String,
location: Location,
source: datatypes::error::Error,
},
#[snafu(display(
"Invalid column default constraint, column: {}, source: {}",
column,
source
))]
#[snafu(display("Invalid column default constraint, column: {}", column))]
InvalidColumnDefaultConstraint {
column: String,
location: Location,

File diff suppressed because it is too large Load Diff

View File

@@ -12,7 +12,9 @@
// See the License for the specific language governing permissions and
// limitations under the License.
use datatypes::schema::{ColumnDefaultConstraint, ColumnSchema};
use std::collections::HashMap;
use datatypes::schema::{ColumnDefaultConstraint, ColumnSchema, COMMENT_KEY};
use snafu::ResultExt;
use crate::error::{self, Result};
@@ -20,7 +22,7 @@ use crate::helper::ColumnDataTypeWrapper;
use crate::v1::ColumnDef;
pub fn try_as_column_schema(column_def: &ColumnDef) -> Result<ColumnSchema> {
let data_type = ColumnDataTypeWrapper::try_new(column_def.datatype)?;
let data_type = ColumnDataTypeWrapper::try_new(column_def.data_type)?;
let constraint = if column_def.default_constraint.is_empty() {
None
@@ -34,9 +36,17 @@ pub fn try_as_column_schema(column_def: &ColumnDef) -> Result<ColumnSchema> {
)
};
ColumnSchema::new(&column_def.name, data_type.into(), column_def.is_nullable)
.with_default_constraint(constraint)
.context(error::InvalidColumnDefaultConstraintSnafu {
column: &column_def.name,
})
let mut metadata = HashMap::new();
if !column_def.comment.is_empty() {
metadata.insert(COMMENT_KEY.to_string(), column_def.comment.clone());
}
Ok(
ColumnSchema::new(&column_def.name, data_type.into(), column_def.is_nullable)
.with_default_constraint(constraint)
.context(error::InvalidColumnDefaultConstraintSnafu {
column: &column_def.name,
})?
.with_metadata(metadata),
)
}

25
src/auth/Cargo.toml Normal file
View File

@@ -0,0 +1,25 @@
[package]
name = "auth"
version.workspace = true
edition.workspace = true
license.workspace = true
[features]
default = []
testing = []
[dependencies]
api.workspace = true
async-trait.workspace = true
common-error.workspace = true
common-macro.workspace = true
digest = "0.10"
hex = { version = "0.4" }
secrecy = { version = "0.8", features = ["serde", "alloc"] }
sha1 = "0.10"
snafu.workspace = true
sql.workspace = true
tokio.workspace = true
[dev-dependencies]
common-test-util.workspace = true

147
src/auth/src/common.rs Normal file
View File

@@ -0,0 +1,147 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::sync::Arc;
use digest::Digest;
use secrecy::SecretString;
use sha1::Sha1;
use snafu::{ensure, OptionExt};
use crate::error::{IllegalParamSnafu, InvalidConfigSnafu, Result, UserPasswordMismatchSnafu};
use crate::user_info::DefaultUserInfo;
use crate::user_provider::static_user_provider::{StaticUserProvider, STATIC_USER_PROVIDER};
use crate::{UserInfoRef, UserProviderRef};
pub(crate) const DEFAULT_USERNAME: &str = "greptime";
/// construct a [`UserInfo`](crate::user_info::UserInfo) impl with name
/// use default username `greptime` if None is provided
pub fn userinfo_by_name(username: Option<String>) -> UserInfoRef {
DefaultUserInfo::with_name(username.unwrap_or_else(|| DEFAULT_USERNAME.to_string()))
}
pub fn user_provider_from_option(opt: &String) -> Result<UserProviderRef> {
let (name, content) = opt.split_once(':').context(InvalidConfigSnafu {
value: opt.to_string(),
msg: "UserProviderOption must be in format `<option>:<value>`",
})?;
match name {
STATIC_USER_PROVIDER => {
let provider =
StaticUserProvider::try_from(content).map(|p| Arc::new(p) as UserProviderRef)?;
Ok(provider)
}
_ => InvalidConfigSnafu {
value: name.to_string(),
msg: "Invalid UserProviderOption",
}
.fail(),
}
}
type Username<'a> = &'a str;
type HostOrIp<'a> = &'a str;
#[derive(Debug, Clone)]
pub enum Identity<'a> {
UserId(Username<'a>, Option<HostOrIp<'a>>),
}
pub type HashedPassword<'a> = &'a [u8];
pub type Salt<'a> = &'a [u8];
/// Authentication information sent by the client.
pub enum Password<'a> {
PlainText(SecretString),
MysqlNativePassword(HashedPassword<'a>, Salt<'a>),
PgMD5(HashedPassword<'a>, Salt<'a>),
}
pub fn auth_mysql(
auth_data: HashedPassword,
salt: Salt,
username: &str,
save_pwd: &[u8],
) -> Result<()> {
ensure!(
auth_data.len() == 20,
IllegalParamSnafu {
msg: "Illegal mysql password length"
}
);
// ref: https://github.com/mysql/mysql-server/blob/a246bad76b9271cb4333634e954040a970222e0a/sql/auth/password.cc#L62
let hash_stage_2 = double_sha1(save_pwd);
let tmp = sha1_two(salt, &hash_stage_2);
// xor auth_data and tmp
let mut xor_result = [0u8; 20];
for i in 0..20 {
xor_result[i] = auth_data[i] ^ tmp[i];
}
let candidate_stage_2 = sha1_one(&xor_result);
if candidate_stage_2 == hash_stage_2 {
Ok(())
} else {
UserPasswordMismatchSnafu {
username: username.to_string(),
}
.fail()
}
}
fn sha1_two(input_1: &[u8], input_2: &[u8]) -> Vec<u8> {
let mut hasher = Sha1::new();
hasher.update(input_1);
hasher.update(input_2);
hasher.finalize().to_vec()
}
fn sha1_one(data: &[u8]) -> Vec<u8> {
let mut hasher = Sha1::new();
hasher.update(data);
hasher.finalize().to_vec()
}
fn double_sha1(data: &[u8]) -> Vec<u8> {
sha1_one(&sha1_one(data))
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_sha() {
let sha_1_answer: Vec<u8> = vec![
124, 74, 141, 9, 202, 55, 98, 175, 97, 229, 149, 32, 148, 61, 194, 100, 148, 248, 148,
27,
];
let sha_1 = sha1_one("123456".as_bytes());
assert_eq!(sha_1, sha_1_answer);
let double_sha1_answer: Vec<u8> = vec![
107, 180, 131, 126, 183, 67, 41, 16, 94, 228, 86, 141, 218, 125, 198, 126, 210, 202,
42, 217,
];
let double_sha1 = double_sha1("123456".as_bytes());
assert_eq!(double_sha1, double_sha1_answer);
let sha1_2_answer: Vec<u8> = vec![
132, 115, 215, 211, 99, 186, 164, 206, 168, 152, 217, 192, 117, 47, 240, 252, 142, 244,
37, 204,
];
let sha1_2 = sha1_two("123456".as_bytes(), "654321".as_bytes());
assert_eq!(sha1_2, sha1_2_answer);
}
}

93
src/auth/src/error.rs Normal file
View File

@@ -0,0 +1,93 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use common_error::ext::{BoxedError, ErrorExt};
use common_error::status_code::StatusCode;
use common_macro::stack_trace_debug;
use snafu::{Location, Snafu};
#[derive(Snafu)]
#[snafu(visibility(pub))]
#[stack_trace_debug]
pub enum Error {
#[snafu(display("Invalid config value: {}, {}", value, msg))]
InvalidConfig { value: String, msg: String },
#[snafu(display("Illegal param: {}", msg))]
IllegalParam { msg: String },
#[snafu(display("Internal state error: {}", msg))]
InternalState { msg: String },
#[snafu(display("IO error"))]
Io {
#[snafu(source)]
error: std::io::Error,
location: Location,
},
#[snafu(display("Auth failed"))]
AuthBackend {
location: Location,
source: BoxedError,
},
#[snafu(display("User not found, username: {}", username))]
UserNotFound { username: String },
#[snafu(display("Unsupported password type: {}", password_type))]
UnsupportedPasswordType { password_type: String },
#[snafu(display("Username and password does not match, username: {}", username))]
UserPasswordMismatch { username: String },
#[snafu(display(
"Access denied for user '{}' to database '{}-{}'",
username,
catalog,
schema
))]
AccessDenied {
catalog: String,
schema: String,
username: String,
},
#[snafu(display("User is not authorized to perform this action"))]
PermissionDenied { location: Location },
}
impl ErrorExt for Error {
fn status_code(&self) -> StatusCode {
match self {
Error::InvalidConfig { .. } => StatusCode::InvalidArguments,
Error::IllegalParam { .. } => StatusCode::InvalidArguments,
Error::InternalState { .. } => StatusCode::Unexpected,
Error::Io { .. } => StatusCode::Internal,
Error::AuthBackend { .. } => StatusCode::Internal,
Error::UserNotFound { .. } => StatusCode::UserNotFound,
Error::UnsupportedPasswordType { .. } => StatusCode::UnsupportedPasswordType,
Error::UserPasswordMismatch { .. } => StatusCode::UserPasswordMismatch,
Error::AccessDenied { .. } => StatusCode::AccessDenied,
Error::PermissionDenied { .. } => StatusCode::PermissionDenied,
}
}
fn as_any(&self) -> &dyn std::any::Any {
self
}
}
pub type Result<T> = std::result::Result<T, Error>;

34
src/auth/src/lib.rs Normal file
View File

@@ -0,0 +1,34 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
mod common;
pub mod error;
mod permission;
mod user_info;
mod user_provider;
#[cfg(feature = "testing")]
pub mod tests;
pub use common::{
auth_mysql, user_provider_from_option, userinfo_by_name, HashedPassword, Identity, Password,
};
pub use permission::{PermissionChecker, PermissionReq, PermissionResp};
pub use user_info::UserInfo;
pub use user_provider::UserProvider;
/// pub type alias
pub type UserInfoRef = std::sync::Arc<dyn UserInfo>;
pub type UserProviderRef = std::sync::Arc<dyn UserProvider>;
pub type PermissionCheckerRef = std::sync::Arc<dyn PermissionChecker>;

View File

@@ -0,0 +1,64 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::fmt::Debug;
use api::v1::greptime_request::Request;
use sql::statements::statement::Statement;
use crate::error::{PermissionDeniedSnafu, Result};
use crate::{PermissionCheckerRef, UserInfoRef};
#[derive(Debug, Clone)]
pub enum PermissionReq<'a> {
GrpcRequest(&'a Request),
SqlStatement(&'a Statement),
PromQuery,
Opentsdb,
LineProtocol,
PromStoreWrite,
PromStoreRead,
Otlp,
}
#[derive(Debug)]
pub enum PermissionResp {
Allow,
Reject,
}
pub trait PermissionChecker: Send + Sync {
fn check_permission(
&self,
user_info: Option<UserInfoRef>,
req: PermissionReq,
) -> Result<PermissionResp>;
}
impl PermissionChecker for Option<&PermissionCheckerRef> {
fn check_permission(
&self,
user_info: Option<UserInfoRef>,
req: PermissionReq,
) -> Result<PermissionResp> {
match self {
Some(checker) => match checker.check_permission(user_info, req) {
Ok(PermissionResp::Reject) => PermissionDeniedSnafu.fail(),
Ok(PermissionResp::Allow) => Ok(PermissionResp::Allow),
Err(e) => Err(e),
},
None => Ok(PermissionResp::Allow),
}
}
}

View File

@@ -11,14 +11,14 @@
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use secrecy::ExposeSecret;
use servers::auth::user_provider::auth_mysql;
use servers::auth::{
AccessDeniedSnafu, Identity, Password, UnsupportedPasswordTypeSnafu, UserNotFoundSnafu,
UserPasswordMismatchSnafu, UserProvider,
use crate::error::{
AccessDeniedSnafu, Result, UnsupportedPasswordTypeSnafu, UserNotFoundSnafu,
UserPasswordMismatchSnafu,
};
use session::context::UserInfo;
use crate::user_info::DefaultUserInfo;
use crate::{auth_mysql, Identity, Password, UserInfoRef, UserProvider};
pub struct DatabaseAuthInfo<'a> {
pub catalog: &'a str,
@@ -56,17 +56,13 @@ impl UserProvider for MockUserProvider {
"mock_user_provider"
}
async fn authenticate(
&self,
id: Identity<'_>,
password: Password<'_>,
) -> servers::auth::Result<UserInfo> {
async fn authenticate(&self, id: Identity<'_>, password: Password<'_>) -> Result<UserInfoRef> {
match id {
Identity::UserId(username, _host) => match password {
Password::PlainText(password) => {
if username == "greptime" {
if password.expose_secret() == "greptime" {
Ok(UserInfo::new("greptime"))
Ok(DefaultUserInfo::with_name("greptime"))
} else {
UserPasswordMismatchSnafu {
username: username.to_string(),
@@ -82,7 +78,7 @@ impl UserProvider for MockUserProvider {
}
Password::MysqlNativePassword(auth_data, salt) => {
auth_mysql(auth_data, salt, username, "greptime".as_bytes())
.map(|_| UserInfo::new(username))
.map(|_| DefaultUserInfo::with_name(username))
}
_ => UnsupportedPasswordTypeSnafu {
password_type: "mysql_native_password",
@@ -92,12 +88,7 @@ impl UserProvider for MockUserProvider {
}
}
async fn authorize(
&self,
catalog: &str,
schema: &str,
user_info: &UserInfo,
) -> servers::auth::Result<()> {
async fn authorize(&self, catalog: &str, schema: &str, user_info: &UserInfoRef) -> Result<()> {
if catalog == self.catalog && schema == self.schema && user_info.username() == self.username
{
Ok(())
@@ -114,6 +105,8 @@ impl UserProvider for MockUserProvider {
#[tokio::test]
async fn test_auth_by_plain_text() {
use crate::error;
let user_provider = MockUserProvider::default();
assert_eq!("mock_user_provider", user_provider.name());
@@ -137,7 +130,7 @@ async fn test_auth_by_plain_text() {
assert!(auth_result.is_err());
assert!(matches!(
auth_result.err().unwrap(),
servers::auth::Error::UnsupportedPasswordType { .. }
error::Error::UnsupportedPasswordType { .. }
));
// auth failed, err: user not exist.
@@ -150,7 +143,7 @@ async fn test_auth_by_plain_text() {
assert!(auth_result.is_err());
assert!(matches!(
auth_result.err().unwrap(),
servers::auth::Error::UserNotFound { .. }
error::Error::UserNotFound { .. }
));
// auth failed, err: wrong password
@@ -163,7 +156,7 @@ async fn test_auth_by_plain_text() {
assert!(auth_result.is_err());
assert!(matches!(
auth_result.err().unwrap(),
servers::auth::Error::UserPasswordMismatch { .. }
error::Error::UserPasswordMismatch { .. }
))
}
@@ -176,8 +169,8 @@ async fn test_schema_validate() {
username: "test_user",
});
let right_user = UserInfo::new("test_user");
let wrong_user = UserInfo::default();
let right_user = DefaultUserInfo::with_name("test_user");
let wrong_user = DefaultUserInfo::with_name("greptime");
// check catalog
let re = validator

47
src/auth/src/user_info.rs Normal file
View File

@@ -0,0 +1,47 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::any::Any;
use std::fmt::Debug;
use std::sync::Arc;
use crate::UserInfoRef;
pub trait UserInfo: Debug + Sync + Send {
fn as_any(&self) -> &dyn Any;
fn username(&self) -> &str;
}
#[derive(Debug)]
pub(crate) struct DefaultUserInfo {
username: String,
}
impl DefaultUserInfo {
pub(crate) fn with_name(username: impl Into<String>) -> UserInfoRef {
Arc::new(Self {
username: username.into(),
})
}
}
impl UserInfo for DefaultUserInfo {
fn as_any(&self) -> &dyn Any {
self
}
fn username(&self) -> &str {
self.username.as_str()
}
}

View File

@@ -0,0 +1,46 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
pub(crate) mod static_user_provider;
use crate::common::{Identity, Password};
use crate::error::Result;
use crate::UserInfoRef;
#[async_trait::async_trait]
pub trait UserProvider: Send + Sync {
fn name(&self) -> &str;
/// Checks whether a user is valid and allowed to access the database.
async fn authenticate(&self, id: Identity<'_>, password: Password<'_>) -> Result<UserInfoRef>;
/// Checks whether a connection request
/// from a certain user to a certain catalog/schema is legal.
/// This method should be called after [authenticate()](UserProvider::authenticate()).
async fn authorize(&self, catalog: &str, schema: &str, user_info: &UserInfoRef) -> Result<()>;
/// Combination of [authenticate()](UserProvider::authenticate()) and [authorize()](UserProvider::authorize()).
/// In most cases it's preferred for both convenience and performance.
async fn auth(
&self,
id: Identity<'_>,
password: Password<'_>,
catalog: &str,
schema: &str,
) -> Result<UserInfoRef> {
let user_info = self.authenticate(id, password).await?;
self.authorize(catalog, schema, &user_info).await?;
Ok(user_info)
}
}

View File

@@ -19,20 +19,17 @@ use std::io::BufRead;
use std::path::Path;
use async_trait::async_trait;
use digest;
use digest::Digest;
use secrecy::ExposeSecret;
use session::context::UserInfo;
use sha1::Sha1;
use snafu::{ensure, OptionExt, ResultExt};
use crate::auth::{
Error, HashedPassword, Identity, IllegalParamSnafu, InvalidConfigSnafu, IoSnafu, Password,
Result, Salt, UnsupportedPasswordTypeSnafu, UserNotFoundSnafu, UserPasswordMismatchSnafu,
UserProvider,
use crate::error::{
Error, IllegalParamSnafu, InvalidConfigSnafu, IoSnafu, Result, UnsupportedPasswordTypeSnafu,
UserNotFoundSnafu, UserPasswordMismatchSnafu,
};
use crate::user_info::DefaultUserInfo;
use crate::{auth_mysql, Identity, Password, UserInfoRef, UserProvider};
pub const STATIC_USER_PROVIDER: &str = "static_user_provider";
pub(crate) const STATIC_USER_PROVIDER: &str = "static_user_provider";
impl TryFrom<&str> for StaticUserProvider {
type Error = Error;
@@ -91,7 +88,7 @@ impl TryFrom<&str> for StaticUserProvider {
}
}
pub struct StaticUserProvider {
pub(crate) struct StaticUserProvider {
users: HashMap<String, Vec<u8>>,
}
@@ -105,7 +102,7 @@ impl UserProvider for StaticUserProvider {
&self,
input_id: Identity<'_>,
input_pwd: Password<'_>,
) -> Result<UserInfo> {
) -> Result<UserInfoRef> {
match input_id {
Identity::UserId(username, _) => {
ensure!(
@@ -127,7 +124,7 @@ impl UserProvider for StaticUserProvider {
}
);
return if save_pwd == pwd.expose_secret().as_bytes() {
Ok(UserInfo::new(username))
Ok(DefaultUserInfo::with_name(username))
} else {
UserPasswordMismatchSnafu {
username: username.to_string(),
@@ -136,14 +133,8 @@ impl UserProvider for StaticUserProvider {
};
}
Password::MysqlNativePassword(auth_data, salt) => {
ensure!(
auth_data.len() == 20,
IllegalParamSnafu {
msg: "Illegal MySQL native password format, length != 20"
}
);
auth_mysql(auth_data, salt, username, save_pwd)
.map(|_| UserInfo::new(username))
.map(|_| DefaultUserInfo::with_name(username))
}
Password::PgMD5(_, _) => UnsupportedPasswordTypeSnafu {
password_type: "pg_md5",
@@ -154,88 +145,28 @@ impl UserProvider for StaticUserProvider {
}
}
async fn authorize(&self, _catalog: &str, _schema: &str, _user_info: &UserInfo) -> Result<()> {
async fn authorize(
&self,
_catalog: &str,
_schema: &str,
_user_info: &UserInfoRef,
) -> Result<()> {
// default allow all
Ok(())
}
}
pub fn auth_mysql(
auth_data: HashedPassword,
salt: Salt,
username: &str,
save_pwd: &[u8],
) -> Result<()> {
// ref: https://github.com/mysql/mysql-server/blob/a246bad76b9271cb4333634e954040a970222e0a/sql/auth/password.cc#L62
let hash_stage_2 = double_sha1(save_pwd);
let tmp = sha1_two(salt, &hash_stage_2);
// xor auth_data and tmp
let mut xor_result = [0u8; 20];
for i in 0..20 {
xor_result[i] = auth_data[i] ^ tmp[i];
}
let candidate_stage_2 = sha1_one(&xor_result);
if candidate_stage_2 == hash_stage_2 {
Ok(())
} else {
UserPasswordMismatchSnafu {
username: username.to_string(),
}
.fail()
}
}
fn sha1_two(input_1: &[u8], input_2: &[u8]) -> Vec<u8> {
let mut hasher = Sha1::new();
hasher.update(input_1);
hasher.update(input_2);
hasher.finalize().to_vec()
}
fn sha1_one(data: &[u8]) -> Vec<u8> {
let mut hasher = Sha1::new();
hasher.update(data);
hasher.finalize().to_vec()
}
fn double_sha1(data: &[u8]) -> Vec<u8> {
sha1_one(&sha1_one(data))
}
#[cfg(test)]
pub mod test {
use std::fs::File;
use std::io::{LineWriter, Write};
use common_test_util::temp_dir::create_temp_dir;
use session::context::UserInfo;
use crate::auth::user_provider::{double_sha1, sha1_one, sha1_two, StaticUserProvider};
use crate::auth::{Identity, Password, UserProvider};
#[test]
fn test_sha() {
let sha_1_answer: Vec<u8> = vec![
124, 74, 141, 9, 202, 55, 98, 175, 97, 229, 149, 32, 148, 61, 194, 100, 148, 248, 148,
27,
];
let sha_1 = sha1_one("123456".as_bytes());
assert_eq!(sha_1, sha_1_answer);
let double_sha1_answer: Vec<u8> = vec![
107, 180, 131, 126, 183, 67, 41, 16, 94, 228, 86, 141, 218, 125, 198, 126, 210, 202,
42, 217,
];
let double_sha1 = double_sha1("123456".as_bytes());
assert_eq!(double_sha1, double_sha1_answer);
let sha1_2_answer: Vec<u8> = vec![
132, 115, 215, 211, 99, 186, 164, 206, 168, 152, 217, 192, 117, 47, 240, 252, 142, 244,
37, 204,
];
let sha1_2 = sha1_two("123456".as_bytes(), "654321".as_bytes());
assert_eq!(sha1_2, sha1_2_answer);
}
use crate::user_info::DefaultUserInfo;
use crate::user_provider::static_user_provider::StaticUserProvider;
use crate::user_provider::{Identity, Password};
use crate::UserProvider;
async fn test_authenticate(provider: &dyn UserProvider, username: &str, password: &str) {
let re = provider
@@ -249,9 +180,10 @@ pub mod test {
#[tokio::test]
async fn test_authorize() {
let user_info = DefaultUserInfo::with_name("root");
let provider = StaticUserProvider::try_from("cmd:root=123456,admin=654321").unwrap();
provider
.authorize("catalog", "schema", &UserInfo::new("root"))
.authorize("catalog", "schema", &user_info)
.await
.unwrap();
}

61
src/auth/tests/mod.rs Normal file
View File

@@ -0,0 +1,61 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#![feature(assert_matches)]
use std::assert_matches::assert_matches;
use std::sync::Arc;
use api::v1::greptime_request::Request;
use auth::error::Error::InternalState;
use auth::{PermissionChecker, PermissionCheckerRef, PermissionReq, PermissionResp, UserInfoRef};
use sql::statements::show::{ShowDatabases, ShowKind};
use sql::statements::statement::Statement;
struct DummyPermissionChecker;
impl PermissionChecker for DummyPermissionChecker {
fn check_permission(
&self,
_user_info: Option<UserInfoRef>,
req: PermissionReq,
) -> auth::error::Result<PermissionResp> {
match req {
PermissionReq::GrpcRequest(_) => Ok(PermissionResp::Allow),
PermissionReq::SqlStatement(_) => Ok(PermissionResp::Reject),
_ => Err(InternalState {
msg: "testing".to_string(),
}),
}
}
}
#[test]
fn test_permission_checker() {
let checker: PermissionCheckerRef = Arc::new(DummyPermissionChecker);
let grpc_result = checker.check_permission(
None,
PermissionReq::GrpcRequest(&Request::Query(Default::default())),
);
assert_matches!(grpc_result, Ok(PermissionResp::Allow));
let sql_result = checker.check_permission(
None,
PermissionReq::SqlStatement(&Statement::ShowDatabases(ShowDatabases::new(ShowKind::All))),
);
assert_matches!(sql_result, Ok(PermissionResp::Reject));
let err_result = checker.check_permission(None, PermissionReq::Opentsdb);
assert_matches!(err_result, Err(InternalState { msg }) if msg == "testing");
}

View File

@@ -16,6 +16,7 @@ async-trait = "0.1"
common-catalog = { workspace = true }
common-error = { workspace = true }
common-grpc = { workspace = true }
common-macro = { workspace = true }
common-meta = { workspace = true }
common-query = { workspace = true }
common-recordbatch = { workspace = true }
@@ -30,8 +31,9 @@ futures-util.workspace = true
lazy_static.workspace = true
meta-client = { workspace = true }
metrics.workspace = true
moka = { version = "0.11", features = ["future"] }
moka = { workspace = true, features = ["future"] }
parking_lot = "0.12"
partition.workspace = true
regex.workspace = true
serde.workspace = true
serde_json = "1.0"
@@ -46,7 +48,6 @@ catalog = { workspace = true, features = ["testing"] }
chrono.workspace = true
common-test-util = { workspace = true }
log-store = { workspace = true }
mito = { workspace = true, features = ["test"] }
object-store = { workspace = true }
storage = { workspace = true }
tokio.workspace = true

View File

@@ -17,41 +17,48 @@ use std::fmt::Debug;
use common_error::ext::{BoxedError, ErrorExt};
use common_error::status_code::StatusCode;
use common_macro::stack_trace_debug;
use datafusion::error::DataFusionError;
use datatypes::prelude::ConcreteDataType;
use snafu::{Location, Snafu};
use table::metadata::TableId;
use tokio::task::JoinError;
use crate::DeregisterTableRequest;
#[derive(Debug, Snafu)]
#[derive(Snafu)]
#[snafu(visibility(pub))]
#[stack_trace_debug]
pub enum Error {
#[snafu(display(
"Failed to re-compile script due to internal error, source: {}",
source
))]
#[snafu(display("Failed to list catalogs"))]
ListCatalogs {
location: Location,
source: BoxedError,
},
#[snafu(display("Failed to list {}'s schemas", catalog))]
ListSchemas {
location: Location,
catalog: String,
source: BoxedError,
},
#[snafu(display("Failed to re-compile script due to internal error"))]
CompileScriptInternal {
location: Location,
source: BoxedError,
},
#[snafu(display("Failed to open system catalog table, source: {}", source))]
#[snafu(display("Failed to open system catalog table"))]
OpenSystemCatalog {
location: Location,
source: table::error::Error,
},
#[snafu(display("Failed to create system catalog table, source: {}", source))]
#[snafu(display("Failed to create system catalog table"))]
CreateSystemCatalog {
location: Location,
source: table::error::Error,
},
#[snafu(display(
"Failed to create table, table info: {}, source: {}",
table_info,
source
))]
#[snafu(display("Failed to create table, table info: {}", table_info))]
CreateTable {
table_info: String,
location: Location,
@@ -85,13 +92,14 @@ pub enum Error {
#[snafu(display("Catalog value is not present"))]
EmptyValue { location: Location },
#[snafu(display("Failed to deserialize value, source: {}", source))]
#[snafu(display("Failed to deserialize value"))]
ValueDeserialize {
source: serde_json::error::Error,
#[snafu(source)]
error: serde_json::error::Error,
location: Location,
},
#[snafu(display("Table engine not found: {}, source: {}", engine_name, source))]
#[snafu(display("Table engine not found: {}", engine_name))]
TableEngineNotFound {
engine_name: String,
location: Location,
@@ -129,15 +137,18 @@ pub enum Error {
#[snafu(display("Operation {} not supported", op))]
NotSupported { op: String, location: Location },
#[snafu(display("Failed to open table, table info: {}, source: {}", table_info, source))]
#[snafu(display("Failed to open table {table_id}"))]
OpenTable {
table_info: String,
table_id: TableId,
location: Location,
source: table::error::Error,
},
#[snafu(display("Failed to open table in parallel, source: {}", source))]
ParallelOpenTable { source: JoinError },
#[snafu(display("Failed to open table in parallel"))]
ParallelOpenTable {
#[snafu(source)]
error: JoinError,
},
#[snafu(display("Table not found while opening table, table info: {}", table_info))]
TableNotFound {
@@ -151,72 +162,52 @@ pub enum Error {
source: common_recordbatch::error::Error,
},
#[snafu(display("Failed to create recordbatch, source: {}", source))]
#[snafu(display("Failed to create recordbatch"))]
CreateRecordBatch {
location: Location,
source: common_recordbatch::error::Error,
},
#[snafu(display(
"Failed to insert table creation record to system catalog, source: {}",
source
))]
#[snafu(display("Failed to insert table creation record to system catalog"))]
InsertCatalogRecord {
location: Location,
source: table::error::Error,
},
#[snafu(display(
"Failed to deregister table, request: {:?}, source: {}",
request,
source
))]
DeregisterTable {
request: DeregisterTableRequest,
location: Location,
source: table::error::Error,
},
#[snafu(display("Illegal catalog manager state: {}", msg))]
IllegalManagerState { location: Location, msg: String },
#[snafu(display("Failed to scan system catalog table, source: {}", source))]
#[snafu(display("Failed to scan system catalog table"))]
SystemCatalogTableScan {
location: Location,
source: table::error::Error,
},
#[snafu(display("{source}"))]
#[snafu(display(""))]
Internal {
location: Location,
source: BoxedError,
},
#[snafu(display(
"Failed to upgrade weak catalog manager reference. location: {}",
location
))]
#[snafu(display("Failed to upgrade weak catalog manager reference"))]
UpgradeWeakCatalogManagerRef { location: Location },
#[snafu(display("Failed to execute system catalog table scan, source: {}", source))]
#[snafu(display("Failed to execute system catalog table scan"))]
SystemCatalogTableScanExec {
location: Location,
source: common_query::error::Error,
},
#[snafu(display("Cannot parse catalog value, source: {}", source))]
#[snafu(display("Cannot parse catalog value"))]
InvalidCatalogValue {
location: Location,
source: common_catalog::error::Error,
},
#[snafu(display("Failed to perform metasrv operation, source: {}", source))]
#[snafu(display("Failed to perform metasrv operation"))]
MetaSrv {
location: Location,
source: meta_client::error::Error,
},
#[snafu(display("Invalid table info in catalog, source: {}", source))]
#[snafu(display("Invalid table info in catalog"))]
InvalidTableInfoInCatalog {
location: Location,
source: datatypes::error::Error,
@@ -225,17 +216,14 @@ pub enum Error {
#[snafu(display("Illegal access to catalog: {} and schema: {}", catalog, schema))]
QueryAccessDenied { catalog: String, schema: String },
#[snafu(display("Invalid system table definition: {err_msg}"))]
InvalidSystemTableDef { err_msg: String, location: Location },
#[snafu(display("{}: {}", msg, source))]
#[snafu(display(""))]
Datafusion {
msg: String,
source: DataFusionError,
#[snafu(source)]
error: DataFusionError,
location: Location,
},
#[snafu(display("Table schema mismatch, source: {}", source))]
#[snafu(display("Table schema mismatch"))]
TableSchemaMismatch {
location: Location,
source: table::error::Error,
@@ -244,7 +232,7 @@ pub enum Error {
#[snafu(display("A generic error has occurred, msg: {}", msg))]
Generic { msg: String, location: Location },
#[snafu(display("Table metadata manager error: {}", source))]
#[snafu(display("Table metadata manager error"))]
TableMetadataManager {
source: common_meta::error::Error,
location: Location,
@@ -259,10 +247,8 @@ impl ErrorExt for Error {
Error::InvalidKey { .. }
| Error::SchemaNotFound { .. }
| Error::TableNotFound { .. }
| Error::IllegalManagerState { .. }
| Error::CatalogNotFound { .. }
| Error::InvalidEntryType { .. }
| Error::InvalidSystemTableDef { .. }
| Error::ParallelOpenTable { .. } => StatusCode::Unexpected,
Error::SystemCatalog { .. }
@@ -284,12 +270,15 @@ impl ErrorExt for Error {
StatusCode::InvalidArguments
}
Error::ListCatalogs { source, .. } | Error::ListSchemas { source, .. } => {
source.status_code()
}
Error::OpenSystemCatalog { source, .. }
| Error::CreateSystemCatalog { source, .. }
| Error::InsertCatalogRecord { source, .. }
| Error::OpenTable { source, .. }
| Error::CreateTable { source, .. }
| Error::DeregisterTable { source, .. }
| Error::TableSchemaMismatch { source, .. } => source.status_code(),
Error::MetaSrv { source, .. } => source.status_code(),

View File

@@ -15,30 +15,27 @@
mod columns;
mod tables;
use std::any::Any;
use std::collections::HashMap;
use std::sync::{Arc, Weak};
use async_trait::async_trait;
use common_catalog::consts::{
INFORMATION_SCHEMA_COLUMNS_TABLE_ID, INFORMATION_SCHEMA_NAME,
INFORMATION_SCHEMA_TABLES_TABLE_ID,
};
use common_catalog::consts::INFORMATION_SCHEMA_NAME;
use common_error::ext::BoxedError;
use common_recordbatch::{RecordBatchStreamAdaptor, SendableRecordBatchStream};
use datatypes::schema::SchemaRef;
use futures_util::StreamExt;
use snafu::ResultExt;
use store_api::data_source::DataSource;
use store_api::storage::{ScanRequest, TableId};
use table::data_source::DataSource;
use table::error::{SchemaConversionSnafu, TablesRecordBatchSnafu};
use table::metadata::{TableIdent, TableInfoBuilder, TableMetaBuilder, TableType};
use table::{Result as TableResult, Table, TableRef};
use table::metadata::{
FilterPushDownType, TableInfoBuilder, TableInfoRef, TableMetaBuilder, TableType,
};
use table::thin_table::{ThinTable, ThinTableAdapter};
use table::TableRef;
use self::columns::InformationSchemaColumns;
use crate::error::Result;
use crate::information_schema::tables::InformationSchemaTables;
use crate::table_factory::TableFactory;
use crate::CatalogManager;
pub const TABLES: &str = "tables";
@@ -63,192 +60,117 @@ impl InformationSchemaProvider {
catalog_name: String,
catalog_manager: Weak<dyn CatalogManager>,
) -> HashMap<String, TableRef> {
let provider = Self::new(catalog_name, catalog_manager);
let mut schema = HashMap::new();
schema.insert(
TABLES.to_string(),
Arc::new(InformationTable::new(
catalog_name.clone(),
INFORMATION_SCHEMA_TABLES_TABLE_ID,
TABLES.to_string(),
Arc::new(InformationSchemaTables::new(
catalog_name.clone(),
catalog_manager.clone(),
)),
)) as _,
);
schema.insert(
COLUMNS.to_string(),
Arc::new(InformationTable::new(
catalog_name.clone(),
INFORMATION_SCHEMA_COLUMNS_TABLE_ID,
COLUMNS.to_string(),
Arc::new(InformationSchemaColumns::new(catalog_name, catalog_manager)),
)) as _,
);
schema.insert(TABLES.to_owned(), provider.table(TABLES).unwrap());
schema.insert(COLUMNS.to_owned(), provider.table(COLUMNS).unwrap());
schema
}
pub fn table(&self, name: &str) -> Result<Option<TableRef>> {
let (stream_builder, table_id) = match name.to_ascii_lowercase().as_ref() {
TABLES => (
Arc::new(InformationSchemaTables::new(
self.catalog_name.clone(),
self.catalog_manager.clone(),
)) as _,
INFORMATION_SCHEMA_TABLES_TABLE_ID,
),
COLUMNS => (
Arc::new(InformationSchemaColumns::new(
self.catalog_name.clone(),
self.catalog_manager.clone(),
)) as _,
INFORMATION_SCHEMA_COLUMNS_TABLE_ID,
),
_ => {
return Ok(None);
}
};
pub fn table(&self, name: &str) -> Option<TableRef> {
self.information_table(name).map(|table| {
let table_info = Self::table_info(self.catalog_name.clone(), &table);
let filter_pushdown = FilterPushDownType::Unsupported;
let thin_table = ThinTable::new(table_info, filter_pushdown);
Ok(Some(Arc::new(InformationTable::new(
self.catalog_name.clone(),
table_id,
name.to_string(),
stream_builder,
))))
let data_source = Arc::new(InformationTableDataSource::new(table));
Arc::new(ThinTableAdapter::new(thin_table, data_source)) as _
})
}
pub fn table_factory(&self, name: &str) -> Result<Option<TableFactory>> {
let (stream_builder, table_id) = match name.to_ascii_lowercase().as_ref() {
TABLES => (
Arc::new(InformationSchemaTables::new(
self.catalog_name.clone(),
self.catalog_manager.clone(),
)) as _,
INFORMATION_SCHEMA_TABLES_TABLE_ID,
),
COLUMNS => (
Arc::new(InformationSchemaColumns::new(
self.catalog_name.clone(),
self.catalog_manager.clone(),
)) as _,
INFORMATION_SCHEMA_COLUMNS_TABLE_ID,
),
_ => {
return Ok(None);
}
};
let data_source = Arc::new(InformationTable::new(
self.catalog_name.clone(),
table_id,
name.to_string(),
stream_builder,
));
Ok(Some(Arc::new(move || data_source.clone())))
}
}
// TODO(ruihang): make it a more generic trait:
// https://github.com/GreptimeTeam/greptimedb/pull/1639#discussion_r1205001903
pub trait InformationStreamBuilder: Send + Sync {
fn to_stream(&self) -> Result<SendableRecordBatchStream>;
fn schema(&self) -> SchemaRef;
}
pub struct InformationTable {
catalog_name: String,
table_id: TableId,
name: String,
stream_builder: Arc<dyn InformationStreamBuilder>,
}
impl InformationTable {
pub fn new(
catalog_name: String,
table_id: TableId,
name: String,
stream_builder: Arc<dyn InformationStreamBuilder>,
) -> Self {
Self {
catalog_name,
table_id,
name,
stream_builder,
fn information_table(&self, name: &str) -> Option<InformationTableRef> {
match name.to_ascii_lowercase().as_str() {
TABLES => Some(Arc::new(InformationSchemaTables::new(
self.catalog_name.clone(),
self.catalog_manager.clone(),
)) as _),
COLUMNS => Some(Arc::new(InformationSchemaColumns::new(
self.catalog_name.clone(),
self.catalog_manager.clone(),
)) as _),
_ => None,
}
}
}
#[async_trait]
impl Table for InformationTable {
fn as_any(&self) -> &dyn Any {
self
}
fn schema(&self) -> SchemaRef {
self.stream_builder.schema()
}
fn table_info(&self) -> table::metadata::TableInfoRef {
fn table_info(catalog_name: String, table: &InformationTableRef) -> TableInfoRef {
let table_meta = TableMetaBuilder::default()
.schema(self.stream_builder.schema())
.schema(table.schema())
.primary_key_indices(vec![])
.next_column_id(0)
.build()
.unwrap();
Arc::new(
TableInfoBuilder::default()
.ident(TableIdent {
table_id: self.table_id,
version: 0,
})
.name(self.name.clone())
.catalog_name(self.catalog_name.clone())
.schema_name(INFORMATION_SCHEMA_NAME.to_string())
.meta(table_meta)
.table_type(TableType::Temporary)
.build()
.unwrap(),
)
let table_info = TableInfoBuilder::default()
.table_id(table.table_id())
.name(table.table_name().to_owned())
.catalog_name(catalog_name)
.schema_name(INFORMATION_SCHEMA_NAME.to_owned())
.meta(table_meta)
.table_type(table.table_type())
.build()
.unwrap();
Arc::new(table_info)
}
}
trait InformationTable {
fn table_id(&self) -> TableId;
fn table_name(&self) -> &'static str;
fn schema(&self) -> SchemaRef;
fn to_stream(&self) -> Result<SendableRecordBatchStream>;
fn table_type(&self) -> TableType {
TableType::Temporary
}
}
async fn scan_to_stream(&self, request: ScanRequest) -> TableResult<SendableRecordBatchStream> {
self.get_stream(request)
type InformationTableRef = Arc<dyn InformationTable + Send + Sync>;
struct InformationTableDataSource {
table: InformationTableRef,
}
impl InformationTableDataSource {
fn new(table: InformationTableRef) -> Self {
Self { table }
}
fn try_project(&self, projection: &[usize]) -> std::result::Result<SchemaRef, BoxedError> {
let schema = self
.table
.schema()
.try_project(projection)
.context(SchemaConversionSnafu)
.map_err(BoxedError::new)?;
Ok(Arc::new(schema))
}
}
impl DataSource for InformationTable {
fn get_stream(&self, request: ScanRequest) -> TableResult<SendableRecordBatchStream> {
impl DataSource for InformationTableDataSource {
fn get_stream(
&self,
request: ScanRequest,
) -> std::result::Result<SendableRecordBatchStream, BoxedError> {
let projection = request.projection;
let projected_schema = if let Some(projection) = &projection {
Arc::new(
self.schema()
.try_project(projection)
.context(SchemaConversionSnafu)?,
)
} else {
self.schema()
let projected_schema = match &projection {
Some(projection) => self.try_project(projection)?,
None => self.table.schema(),
};
let stream = self
.stream_builder
.table
.to_stream()
.map_err(BoxedError::new)
.context(TablesRecordBatchSnafu)?
.map(move |batch| {
batch.and_then(|batch| {
if let Some(projection) = &projection {
batch.try_project(projection)
} else {
Ok(batch)
}
})
.context(TablesRecordBatchSnafu)
.map_err(BoxedError::new)?
.map(move |batch| match &projection {
Some(p) => batch.and_then(|b| b.try_project(p)),
None => batch,
});
let stream = RecordBatchStreamAdaptor {
schema: projected_schema,
stream: Box::pin(stream),

View File

@@ -16,8 +16,8 @@ use std::sync::{Arc, Weak};
use arrow_schema::SchemaRef as ArrowSchemaRef;
use common_catalog::consts::{
INFORMATION_SCHEMA_NAME, SEMANTIC_TYPE_FIELD, SEMANTIC_TYPE_PRIMARY_KEY,
SEMANTIC_TYPE_TIME_INDEX,
INFORMATION_SCHEMA_COLUMNS_TABLE_ID, INFORMATION_SCHEMA_NAME, SEMANTIC_TYPE_FIELD,
SEMANTIC_TYPE_PRIMARY_KEY, SEMANTIC_TYPE_TIME_INDEX,
};
use common_error::ext::BoxedError;
use common_query::physical_plan::TaskContext;
@@ -31,9 +31,10 @@ use datatypes::scalars::ScalarVectorBuilder;
use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
use datatypes::vectors::{StringVectorBuilder, VectorRef};
use snafu::{OptionExt, ResultExt};
use store_api::storage::TableId;
use super::tables::InformationSchemaTables;
use super::{InformationStreamBuilder, COLUMNS, TABLES};
use super::{InformationTable, COLUMNS, TABLES};
use crate::error::{
CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
};
@@ -81,7 +82,15 @@ impl InformationSchemaColumns {
}
}
impl InformationStreamBuilder for InformationSchemaColumns {
impl InformationTable for InformationSchemaColumns {
fn table_id(&self) -> TableId {
INFORMATION_SCHEMA_COLUMNS_TABLE_ID
}
fn table_name(&self) -> &'static str {
COLUMNS
}
fn schema(&self) -> SchemaRef {
self.schema.clone()
}
@@ -149,7 +158,7 @@ impl InformationSchemaColumnsBuilder {
for schema_name in catalog_manager.schema_names(&catalog_name).await? {
if !catalog_manager
.schema_exist(&catalog_name, &schema_name)
.schema_exists(&catalog_name, &schema_name)
.await?
{
continue;

View File

@@ -30,13 +30,14 @@ use datatypes::prelude::{ConcreteDataType, ScalarVectorBuilder, VectorRef};
use datatypes::schema::{ColumnSchema, Schema, SchemaRef};
use datatypes::vectors::{StringVectorBuilder, UInt32VectorBuilder};
use snafu::{OptionExt, ResultExt};
use store_api::storage::TableId;
use table::metadata::TableType;
use super::{COLUMNS, TABLES};
use crate::error::{
CreateRecordBatchSnafu, InternalSnafu, Result, UpgradeWeakCatalogManagerRefSnafu,
};
use crate::information_schema::InformationStreamBuilder;
use crate::information_schema::InformationTable;
use crate::CatalogManager;
pub(super) struct InformationSchemaTables {
@@ -74,7 +75,15 @@ impl InformationSchemaTables {
}
}
impl InformationStreamBuilder for InformationSchemaTables {
impl InformationTable for InformationSchemaTables {
fn table_id(&self) -> TableId {
INFORMATION_SCHEMA_TABLES_TABLE_ID
}
fn table_name(&self) -> &'static str {
TABLES
}
fn schema(&self) -> SchemaRef {
self.schema.clone()
}
@@ -145,7 +154,7 @@ impl InformationSchemaTablesBuilder {
for schema_name in catalog_manager.schema_names(&catalog_name).await? {
if !catalog_manager
.schema_exist(&catalog_name, &schema_name)
.schema_exists(&catalog_name, &schema_name)
.await?
{
continue;

View File

@@ -12,11 +12,11 @@
// See the License for the specific language governing permissions and
// limitations under the License.
pub mod immutable;
mod procedure;
#[cfg(test)]
mod tests;
pub use client::{CachedMetaKvBackend, MetaKvBackend};
use table::metadata::TableVersion;
mod client;
mod manager;
const INIT_TABLE_VERSION: TableVersion = 0;
#[cfg(feature = "testing")]
pub mod mock;
pub use manager::KvBackendCatalogManager;

View File

@@ -18,8 +18,9 @@ use std::sync::Arc;
use std::time::Duration;
use common_error::ext::BoxedError;
use common_meta::cache_invalidator::KvCacheInvalidator;
use common_meta::error::Error::{CacheNotGet, GetKvCache};
use common_meta::error::{CacheNotGetSnafu, Error, MetaSrvSnafu, Result};
use common_meta::error::{CacheNotGetSnafu, Error, ExternalSnafu, Result};
use common_meta::kv_backend::{KvBackend, KvBackendRef, TxnService};
use common_meta::rpc::store::{
BatchDeleteRequest, BatchDeleteResponse, BatchGetRequest, BatchGetResponse, BatchPutRequest,
@@ -28,12 +29,11 @@ use common_meta::rpc::store::{
RangeRequest, RangeResponse,
};
use common_meta::rpc::KeyValue;
use common_telemetry::timer;
use common_telemetry::{debug, timer};
use meta_client::client::MetaClient;
use moka::future::{Cache, CacheBuilder};
use snafu::{OptionExt, ResultExt};
use super::KvCacheInvalidator;
use crate::metrics::{METRIC_CATALOG_KV_GET, METRIC_CATALOG_KV_REMOTE_GET};
const CACHE_MAX_CAPACITY: u64 = 10000;
@@ -197,7 +197,8 @@ impl KvBackend for CachedMetaKvBackend {
#[async_trait::async_trait]
impl KvCacheInvalidator for CachedMetaKvBackend {
async fn invalidate_key(&self, key: &[u8]) {
self.cache.invalidate(key).await
self.cache.invalidate(key).await;
debug!("invalidated cache key: {}", String::from_utf8_lossy(key));
}
}
@@ -251,7 +252,7 @@ impl KvBackend for MetaKvBackend {
.range(req)
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)
.context(ExternalSnafu)
}
async fn get(&self, key: &[u8]) -> Result<Option<KeyValue>> {
@@ -260,7 +261,7 @@ impl KvBackend for MetaKvBackend {
.range(RangeRequest::new().with_key(key))
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)?;
.context(ExternalSnafu)?;
Ok(response.take_kvs().get_mut(0).map(|kv| KeyValue {
key: kv.take_key(),
value: kv.take_value(),
@@ -272,7 +273,7 @@ impl KvBackend for MetaKvBackend {
.batch_put(req)
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)
.context(ExternalSnafu)
}
async fn put(&self, req: PutRequest) -> Result<PutResponse> {
@@ -280,7 +281,7 @@ impl KvBackend for MetaKvBackend {
.put(req)
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)
.context(ExternalSnafu)
}
async fn delete_range(&self, req: DeleteRangeRequest) -> Result<DeleteRangeResponse> {
@@ -288,7 +289,7 @@ impl KvBackend for MetaKvBackend {
.delete_range(req)
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)
.context(ExternalSnafu)
}
async fn batch_delete(&self, req: BatchDeleteRequest) -> Result<BatchDeleteResponse> {
@@ -296,7 +297,7 @@ impl KvBackend for MetaKvBackend {
.batch_delete(req)
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)
.context(ExternalSnafu)
}
async fn batch_get(&self, req: BatchGetRequest) -> Result<BatchGetResponse> {
@@ -304,7 +305,7 @@ impl KvBackend for MetaKvBackend {
.batch_get(req)
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)
.context(ExternalSnafu)
}
async fn compare_and_put(
@@ -315,7 +316,7 @@ impl KvBackend for MetaKvBackend {
.compare_and_put(request)
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)
.context(ExternalSnafu)
}
async fn move_value(&self, req: MoveValueRequest) -> Result<MoveValueResponse> {
@@ -323,7 +324,7 @@ impl KvBackend for MetaKvBackend {
.move_value(req)
.await
.map_err(BoxedError::new)
.context(MetaSrvSnafu)
.context(ExternalSnafu)
}
fn as_any(&self) -> &dyn Any {

View File

@@ -0,0 +1,292 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::any::Any;
use std::collections::BTreeSet;
use std::sync::{Arc, Weak};
use common_catalog::consts::{DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME, NUMBERS_TABLE_ID};
use common_error::ext::BoxedError;
use common_meta::cache_invalidator::{CacheInvalidator, CacheInvalidatorRef, Context};
use common_meta::datanode_manager::DatanodeManagerRef;
use common_meta::error::Result as MetaResult;
use common_meta::key::catalog_name::CatalogNameKey;
use common_meta::key::schema_name::SchemaNameKey;
use common_meta::key::table_name::TableNameKey;
use common_meta::key::{TableMetadataManager, TableMetadataManagerRef};
use common_meta::kv_backend::KvBackendRef;
use common_meta::table_name::TableName;
use futures_util::TryStreamExt;
use partition::manager::{PartitionRuleManager, PartitionRuleManagerRef};
use snafu::prelude::*;
use table::dist_table::DistTable;
use table::metadata::TableId;
use table::table::numbers::{NumbersTable, NUMBERS_TABLE_NAME};
use table::TableRef;
use crate::error::{
self as catalog_err, ListCatalogsSnafu, ListSchemasSnafu, Result as CatalogResult,
TableMetadataManagerSnafu,
};
use crate::information_schema::{InformationSchemaProvider, COLUMNS, TABLES};
use crate::CatalogManager;
/// Access all existing catalog, schema and tables.
///
/// The result comes from two source, all the user tables are presented in
/// a kv-backend which persists the metadata of a table. And system tables
/// comes from `SystemCatalog`, which is static and read-only.
#[derive(Clone)]
pub struct KvBackendCatalogManager {
// TODO(LFC): Maybe use a real implementation for Standalone mode.
// Now we use `NoopKvCacheInvalidator` for Standalone mode. In Standalone mode, the KV backend
// is implemented by RaftEngine. Maybe we need a cache for it?
cache_invalidator: CacheInvalidatorRef,
partition_manager: PartitionRuleManagerRef,
table_metadata_manager: TableMetadataManagerRef,
datanode_manager: DatanodeManagerRef,
/// A sub-CatalogManager that handles system tables
system_catalog: SystemCatalog,
}
#[async_trait::async_trait]
impl CacheInvalidator for KvBackendCatalogManager {
async fn invalidate_table_name(&self, ctx: &Context, table_name: TableName) -> MetaResult<()> {
self.cache_invalidator
.invalidate_table_name(ctx, table_name)
.await
}
async fn invalidate_table_id(&self, ctx: &Context, table_id: TableId) -> MetaResult<()> {
self.cache_invalidator
.invalidate_table_id(ctx, table_id)
.await
}
}
impl KvBackendCatalogManager {
pub fn new(
backend: KvBackendRef,
cache_invalidator: CacheInvalidatorRef,
datanode_manager: DatanodeManagerRef,
) -> Arc<Self> {
Arc::new_cyclic(|me| Self {
partition_manager: Arc::new(PartitionRuleManager::new(backend.clone())),
table_metadata_manager: Arc::new(TableMetadataManager::new(backend)),
cache_invalidator,
datanode_manager,
system_catalog: SystemCatalog {
catalog_manager: me.clone(),
},
})
}
pub fn partition_manager(&self) -> PartitionRuleManagerRef {
self.partition_manager.clone()
}
pub fn table_metadata_manager_ref(&self) -> &TableMetadataManagerRef {
&self.table_metadata_manager
}
pub fn datanode_manager(&self) -> DatanodeManagerRef {
self.datanode_manager.clone()
}
}
#[async_trait::async_trait]
impl CatalogManager for KvBackendCatalogManager {
async fn catalog_names(&self) -> CatalogResult<Vec<String>> {
let stream = self
.table_metadata_manager
.catalog_manager()
.catalog_names()
.await;
let keys = stream
.try_collect::<Vec<_>>()
.await
.map_err(BoxedError::new)
.context(ListCatalogsSnafu)?;
Ok(keys)
}
async fn schema_names(&self, catalog: &str) -> CatalogResult<Vec<String>> {
let stream = self
.table_metadata_manager
.schema_manager()
.schema_names(catalog)
.await;
let mut keys = stream
.try_collect::<BTreeSet<_>>()
.await
.map_err(BoxedError::new)
.context(ListSchemasSnafu { catalog })?
.into_iter()
.collect::<Vec<_>>();
keys.extend_from_slice(&self.system_catalog.schema_names());
Ok(keys)
}
async fn table_names(&self, catalog: &str, schema: &str) -> CatalogResult<Vec<String>> {
let mut tables = self
.table_metadata_manager
.table_name_manager()
.tables(catalog, schema)
.await
.context(TableMetadataManagerSnafu)?
.into_iter()
.map(|(k, _)| k)
.collect::<Vec<String>>();
tables.extend_from_slice(&self.system_catalog.table_names(schema));
Ok(tables)
}
async fn catalog_exists(&self, catalog: &str) -> CatalogResult<bool> {
self.table_metadata_manager
.catalog_manager()
.exists(CatalogNameKey::new(catalog))
.await
.context(TableMetadataManagerSnafu)
}
async fn schema_exists(&self, catalog: &str, schema: &str) -> CatalogResult<bool> {
if self.system_catalog.schema_exist(schema) {
return Ok(true);
}
self.table_metadata_manager
.schema_manager()
.exists(SchemaNameKey::new(catalog, schema))
.await
.context(TableMetadataManagerSnafu)
}
async fn table_exists(&self, catalog: &str, schema: &str, table: &str) -> CatalogResult<bool> {
if self.system_catalog.table_exist(schema, table) {
return Ok(true);
}
let key = TableNameKey::new(catalog, schema, table);
self.table_metadata_manager
.table_name_manager()
.get(key)
.await
.context(TableMetadataManagerSnafu)
.map(|x| x.is_some())
}
async fn table(
&self,
catalog: &str,
schema: &str,
table_name: &str,
) -> CatalogResult<Option<TableRef>> {
if let Some(table) = self.system_catalog.table(catalog, schema, table_name) {
return Ok(Some(table));
}
let key = TableNameKey::new(catalog, schema, table_name);
let Some(table_name_value) = self
.table_metadata_manager
.table_name_manager()
.get(key)
.await
.context(TableMetadataManagerSnafu)?
else {
return Ok(None);
};
let table_id = table_name_value.table_id();
let Some(table_info_value) = self
.table_metadata_manager
.table_info_manager()
.get(table_id)
.await
.context(TableMetadataManagerSnafu)?
.map(|v| v.into_inner())
else {
return Ok(None);
};
let table_info = Arc::new(
table_info_value
.table_info
.try_into()
.context(catalog_err::InvalidTableInfoInCatalogSnafu)?,
);
Ok(Some(DistTable::table(table_info)))
}
fn as_any(&self) -> &dyn Any {
self
}
}
// TODO: This struct can hold a static map of all system tables when
// the upper layer (e.g., procedure) can inform the catalog manager
// a new catalog is created.
/// Existing system tables:
/// - public.numbers
/// - information_schema.tables
/// - information_schema.columns
#[derive(Clone)]
struct SystemCatalog {
catalog_manager: Weak<KvBackendCatalogManager>,
}
impl SystemCatalog {
fn schema_names(&self) -> Vec<String> {
vec![INFORMATION_SCHEMA_NAME.to_string()]
}
fn table_names(&self, schema: &str) -> Vec<String> {
if schema == INFORMATION_SCHEMA_NAME {
vec![TABLES.to_string(), COLUMNS.to_string()]
} else if schema == DEFAULT_SCHEMA_NAME {
vec![NUMBERS_TABLE_NAME.to_string()]
} else {
vec![]
}
}
fn schema_exist(&self, schema: &str) -> bool {
schema == INFORMATION_SCHEMA_NAME
}
fn table_exist(&self, schema: &str, table: &str) -> bool {
if schema == INFORMATION_SCHEMA_NAME {
table == TABLES || table == COLUMNS
} else if schema == DEFAULT_SCHEMA_NAME {
table == NUMBERS_TABLE_NAME
} else {
false
}
}
fn table(&self, catalog: &str, schema: &str, table_name: &str) -> Option<TableRef> {
if schema == INFORMATION_SCHEMA_NAME {
let information_schema_provider =
InformationSchemaProvider::new(catalog.to_string(), self.catalog_manager.clone());
information_schema_provider.table(table_name)
} else if schema == DEFAULT_SCHEMA_NAME && table_name == NUMBERS_TABLE_NAME {
Some(NumbersTable::table(NUMBERS_TABLE_ID))
} else {
None
}
}
}

View File

@@ -55,14 +55,14 @@ impl TableEngine for MockTableEngine {
let data = vec![Arc::new(StringVector::from(vec!["a", "b", "c"])) as _];
let record_batch = RecordBatch::new(schema, data).unwrap();
let table: TableRef = Arc::new(MemTable::new_with_catalog(
let table = MemTable::new_with_catalog(
&request.table_name,
record_batch,
table_id,
request.catalog_name,
request.schema_name,
vec![0],
)) as Arc<_>;
);
let mut tables = self.tables.write().unwrap();
let _ = tables.insert(table_id, table.clone() as TableRef);

View File

@@ -17,80 +17,38 @@
#![feature(try_blocks)]
use std::any::Any;
use std::collections::HashMap;
use std::fmt::{Debug, Formatter};
use std::sync::Arc;
use api::v1::meta::{RegionStat, TableIdent, TableName};
use common_telemetry::{info, warn};
use snafu::ResultExt;
use table::engine::{EngineContext, TableEngineRef};
use futures::future::BoxFuture;
use table::metadata::TableId;
use table::requests::CreateTableRequest;
use table::TableRef;
use crate::error::{CreateTableSnafu, Result};
use crate::error::Result;
pub mod error;
pub mod information_schema;
pub mod local;
pub mod kvbackend;
pub mod memory;
mod metrics;
pub mod remote;
pub mod system;
pub mod table_factory;
pub mod table_source;
pub mod tables;
#[async_trait::async_trait]
pub trait CatalogManager: Send + Sync {
fn as_any(&self) -> &dyn Any;
/// Starts a catalog manager.
async fn start(&self) -> Result<()>;
/// Registers a catalog to catalog manager, returns whether the catalog exist before.
async fn register_catalog(self: Arc<Self>, name: String) -> Result<bool>;
/// Register a schema with catalog name and schema name. Retuens whether the
/// schema registered.
///
/// # Errors
///
/// This method will/should fail if catalog not exist
async fn register_schema(&self, request: RegisterSchemaRequest) -> Result<bool>;
/// Deregisters a database within given catalog/schema to catalog manager
async fn deregister_schema(&self, request: DeregisterSchemaRequest) -> Result<bool>;
/// Registers a table within given catalog/schema to catalog manager,
/// returns whether the table registered.
///
/// # Errors
///
/// This method will/should fail if catalog or schema not exist
async fn register_table(&self, request: RegisterTableRequest) -> Result<bool>;
/// Deregisters a table within given catalog/schema to catalog manager
async fn deregister_table(&self, request: DeregisterTableRequest) -> Result<()>;
/// Rename a table to [RenameTableRequest::new_table_name], returns whether the table is renamed.
async fn rename_table(&self, request: RenameTableRequest) -> Result<bool>;
/// Register a system table, should be called before starting the manager.
async fn register_system_table(&self, request: RegisterSystemTableRequest)
-> error::Result<()>;
async fn catalog_names(&self) -> Result<Vec<String>>;
async fn schema_names(&self, catalog: &str) -> Result<Vec<String>>;
async fn table_names(&self, catalog: &str, schema: &str) -> Result<Vec<String>>;
async fn catalog_exist(&self, catalog: &str) -> Result<bool>;
async fn catalog_exists(&self, catalog: &str) -> Result<bool>;
async fn schema_exist(&self, catalog: &str, schema: &str) -> Result<bool>;
async fn schema_exists(&self, catalog: &str, schema: &str) -> Result<bool>;
async fn table_exist(&self, catalog: &str, schema: &str, table: &str) -> Result<bool>;
async fn table_exists(&self, catalog: &str, schema: &str, table: &str) -> Result<bool>;
/// Returns the table by catalog, schema and table name.
async fn table(
@@ -104,7 +62,8 @@ pub trait CatalogManager: Send + Sync {
pub type CatalogManagerRef = Arc<dyn CatalogManager>;
/// Hook called after system table opening.
pub type OpenSystemTableHook = Arc<dyn Fn(TableRef) -> Result<()> + Send + Sync>;
pub type OpenSystemTableHook =
Box<dyn Fn(TableRef) -> BoxFuture<'static, Result<()>> + Send + Sync>;
/// Register system table request:
/// - When system table is already created and registered, the hook will be called
@@ -163,116 +122,3 @@ pub struct RegisterSchemaRequest {
pub catalog: String,
pub schema: String,
}
pub(crate) async fn handle_system_table_request<'a, M: CatalogManager>(
manager: &'a M,
engine: TableEngineRef,
sys_table_requests: &'a mut Vec<RegisterSystemTableRequest>,
) -> Result<()> {
for req in sys_table_requests.drain(..) {
let catalog_name = &req.create_table_request.catalog_name;
let schema_name = &req.create_table_request.schema_name;
let table_name = &req.create_table_request.table_name;
let table_id = req.create_table_request.id;
let table = manager.table(catalog_name, schema_name, table_name).await?;
let table = if let Some(table) = table {
table
} else {
let table = engine
.create_table(&EngineContext::default(), req.create_table_request.clone())
.await
.with_context(|_| CreateTableSnafu {
table_info: common_catalog::format_full_table_name(
catalog_name,
schema_name,
table_name,
),
})?;
let _ = manager
.register_table(RegisterTableRequest {
catalog: catalog_name.clone(),
schema: schema_name.clone(),
table_name: table_name.clone(),
table_id,
table: table.clone(),
})
.await?;
info!("Created and registered system table: {table_name}");
table
};
if let Some(hook) = req.open_hook {
(hook)(table)?;
}
}
Ok(())
}
/// The stat of regions in the datanode node.
/// The number of regions can be got from len of vec.
///
/// Ignores any errors occurred during iterating regions. The intention of this method is to
/// collect region stats that will be carried in Datanode's heartbeat to Metasrv, so it's a
/// "try our best" job.
pub async fn datanode_stat(catalog_manager: &CatalogManagerRef) -> (u64, Vec<RegionStat>) {
let mut region_number: u64 = 0;
let mut region_stats = Vec::new();
let Ok(catalog_names) = catalog_manager.catalog_names().await else {
return (region_number, region_stats);
};
for catalog_name in catalog_names {
let Ok(schema_names) = catalog_manager.schema_names(&catalog_name).await else {
continue;
};
for schema_name in schema_names {
let Ok(table_names) = catalog_manager
.table_names(&catalog_name, &schema_name)
.await
else {
continue;
};
for table_name in table_names {
let Ok(Some(table)) = catalog_manager
.table(&catalog_name, &schema_name, &table_name)
.await
else {
continue;
};
let table_info = table.table_info();
let region_numbers = &table_info.meta.region_numbers;
region_number += region_numbers.len() as u64;
let engine = &table_info.meta.engine;
let table_id = table_info.ident.table_id;
match table.region_stats() {
Ok(stats) => {
let stats = stats.into_iter().map(|stat| RegionStat {
region_id: stat.region_id,
table_ident: Some(TableIdent {
table_id,
table_name: Some(TableName {
catalog_name: catalog_name.clone(),
schema_name: schema_name.clone(),
table_name: table_name.clone(),
}),
engine: engine.clone(),
}),
approximate_bytes: stat.disk_usage_bytes as i64,
attrs: HashMap::from([("engine_name".to_owned(), engine.clone())]),
..Default::default()
});
region_stats.extend(stats);
}
Err(e) => {
warn!("Failed to get region status, err: {:?}", e);
}
};
}
}
}
(region_number, region_stats)
}

View File

@@ -1,634 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::any::Any;
use std::sync::atomic::{AtomicU32, Ordering};
use std::sync::Arc;
use common_catalog::consts::{
DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME, MIN_USER_TABLE_ID,
MITO_ENGINE, NUMBERS_TABLE_ID, SYSTEM_CATALOG_NAME, SYSTEM_CATALOG_TABLE_ID,
SYSTEM_CATALOG_TABLE_NAME,
};
use common_catalog::format_full_table_name;
use common_recordbatch::{RecordBatch, SendableRecordBatchStream};
use common_telemetry::{error, info};
use datatypes::prelude::ScalarVector;
use datatypes::vectors::{BinaryVector, UInt8Vector};
use futures_util::lock::Mutex;
use metrics::increment_gauge;
use snafu::{ensure, OptionExt, ResultExt};
use table::engine::manager::TableEngineManagerRef;
use table::engine::EngineContext;
use table::metadata::TableId;
use table::requests::OpenTableRequest;
use table::table::numbers::{NumbersTable, NUMBERS_TABLE_NAME};
use table::table::TableIdProvider;
use table::TableRef;
use crate::error::{
self, CatalogNotFoundSnafu, IllegalManagerStateSnafu, OpenTableSnafu, ReadSystemCatalogSnafu,
Result, SchemaExistsSnafu, SchemaNotFoundSnafu, SystemCatalogSnafu,
SystemCatalogTypeMismatchSnafu, TableEngineNotFoundSnafu, TableExistsSnafu, TableNotExistSnafu,
TableNotFoundSnafu, UnimplementedSnafu,
};
use crate::local::memory::MemoryCatalogManager;
use crate::system::{
decode_system_catalog, Entry, SystemCatalogTable, TableEntry, ENTRY_TYPE_INDEX, KEY_INDEX,
VALUE_INDEX,
};
use crate::tables::SystemCatalog;
use crate::{
handle_system_table_request, CatalogManager, DeregisterSchemaRequest, DeregisterTableRequest,
RegisterSchemaRequest, RegisterSystemTableRequest, RegisterTableRequest, RenameTableRequest,
};
/// A `CatalogManager` consists of a system catalog and a bunch of user catalogs.
pub struct LocalCatalogManager {
system: Arc<SystemCatalog>,
catalogs: Arc<MemoryCatalogManager>,
engine_manager: TableEngineManagerRef,
next_table_id: AtomicU32,
init_lock: Mutex<bool>,
register_lock: Mutex<()>,
system_table_requests: Mutex<Vec<RegisterSystemTableRequest>>,
}
impl LocalCatalogManager {
/// Create a new [CatalogManager] with given user catalogs and mito engine
pub async fn try_new(engine_manager: TableEngineManagerRef) -> Result<Self> {
let engine = engine_manager
.engine(MITO_ENGINE)
.context(TableEngineNotFoundSnafu {
engine_name: MITO_ENGINE,
})?;
let table = SystemCatalogTable::new(engine.clone()).await?;
let memory_catalog_manager = crate::local::memory::new_memory_catalog_manager()?;
let system_catalog = Arc::new(SystemCatalog::new(table));
Ok(Self {
system: system_catalog,
catalogs: memory_catalog_manager,
engine_manager,
next_table_id: AtomicU32::new(MIN_USER_TABLE_ID),
init_lock: Mutex::new(false),
register_lock: Mutex::new(()),
system_table_requests: Mutex::new(Vec::default()),
})
}
/// Scan all entries from system catalog table
pub async fn init(&self) -> Result<()> {
self.init_system_catalog().await?;
let system_records = self.system.information_schema.system.records().await?;
let entries = self.collect_system_catalog_entries(system_records).await?;
let max_table_id = self.handle_system_catalog_entries(entries).await?;
info!(
"All system catalog entries processed, max table id: {}",
max_table_id
);
self.next_table_id
.store((max_table_id + 1).max(MIN_USER_TABLE_ID), Ordering::Relaxed);
*self.init_lock.lock().await = true;
// Processing system table hooks
let mut sys_table_requests = self.system_table_requests.lock().await;
let engine = self
.engine_manager
.engine(MITO_ENGINE)
.context(TableEngineNotFoundSnafu {
engine_name: MITO_ENGINE,
})?;
handle_system_table_request(self, engine, &mut sys_table_requests).await?;
Ok(())
}
async fn init_system_catalog(&self) -> Result<()> {
// register default catalog and default schema
self.catalogs
.register_catalog_sync(DEFAULT_CATALOG_NAME.to_string())?;
self.catalogs.register_schema_sync(RegisterSchemaRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
})?;
// register SystemCatalogTable
self.catalogs
.register_catalog_sync(SYSTEM_CATALOG_NAME.to_string())?;
self.catalogs.register_schema_sync(RegisterSchemaRequest {
catalog: SYSTEM_CATALOG_NAME.to_string(),
schema: INFORMATION_SCHEMA_NAME.to_string(),
})?;
let register_table_req = RegisterTableRequest {
catalog: SYSTEM_CATALOG_NAME.to_string(),
schema: INFORMATION_SCHEMA_NAME.to_string(),
table_name: SYSTEM_CATALOG_TABLE_NAME.to_string(),
table_id: SYSTEM_CATALOG_TABLE_ID,
table: self.system.information_schema.system.clone(),
};
self.catalogs.register_table(register_table_req).await?;
// Add numbers table for test
let numbers_table = Arc::new(NumbersTable::default());
let register_number_table_req = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: NUMBERS_TABLE_NAME.to_string(),
table_id: NUMBERS_TABLE_ID,
table: numbers_table,
};
self.catalogs
.register_table(register_number_table_req)
.await?;
Ok(())
}
/// Collect stream of system catalog entries to `Vec<Entry>`
async fn collect_system_catalog_entries(
&self,
stream: SendableRecordBatchStream,
) -> Result<Vec<Entry>> {
let record_batch = common_recordbatch::util::collect(stream)
.await
.context(ReadSystemCatalogSnafu)?;
let rbs = record_batch
.into_iter()
.map(Self::record_batch_to_entry)
.collect::<Result<Vec<_>>>()?;
Ok(rbs.into_iter().flat_map(Vec::into_iter).collect::<_>())
}
/// Convert `RecordBatch` to a vector of `Entry`.
fn record_batch_to_entry(rb: RecordBatch) -> Result<Vec<Entry>> {
ensure!(
rb.num_columns() >= 6,
SystemCatalogSnafu {
msg: format!("Length mismatch: {}", rb.num_columns())
}
);
let entry_type = rb
.column(ENTRY_TYPE_INDEX)
.as_any()
.downcast_ref::<UInt8Vector>()
.with_context(|| SystemCatalogTypeMismatchSnafu {
data_type: rb.column(ENTRY_TYPE_INDEX).data_type(),
})?;
let key = rb
.column(KEY_INDEX)
.as_any()
.downcast_ref::<BinaryVector>()
.with_context(|| SystemCatalogTypeMismatchSnafu {
data_type: rb.column(KEY_INDEX).data_type(),
})?;
let value = rb
.column(VALUE_INDEX)
.as_any()
.downcast_ref::<BinaryVector>()
.with_context(|| SystemCatalogTypeMismatchSnafu {
data_type: rb.column(VALUE_INDEX).data_type(),
})?;
let mut res = Vec::with_capacity(rb.num_rows());
for ((t, k), v) in entry_type
.iter_data()
.zip(key.iter_data())
.zip(value.iter_data())
{
let entry = decode_system_catalog(t, k, v)?;
res.push(entry);
}
Ok(res)
}
/// Processes records from system catalog table and returns the max table id persisted
/// in system catalog table.
async fn handle_system_catalog_entries(&self, entries: Vec<Entry>) -> Result<TableId> {
let entries = Self::sort_entries(entries);
let mut max_table_id = 0;
for entry in entries {
match entry {
Entry::Catalog(c) => {
self.catalogs
.register_catalog_sync(c.catalog_name.clone())?;
info!("Register catalog: {}", c.catalog_name);
}
Entry::Schema(s) => {
let req = RegisterSchemaRequest {
catalog: s.catalog_name.clone(),
schema: s.schema_name.clone(),
};
let _ = self.catalogs.register_schema_sync(req)?;
info!("Registered schema: {:?}", s);
}
Entry::Table(t) => {
max_table_id = max_table_id.max(t.table_id);
if t.is_deleted {
continue;
}
self.open_and_register_table(&t).await?;
info!("Registered table: {:?}", t);
}
}
}
Ok(max_table_id)
}
/// Sort catalog entries to ensure catalog entries comes first, then schema entries,
/// and table entries is the last.
fn sort_entries(mut entries: Vec<Entry>) -> Vec<Entry> {
entries.sort();
entries
}
async fn open_and_register_table(&self, t: &TableEntry) -> Result<()> {
self.check_catalog_schema_exist(&t.catalog_name, &t.schema_name)
.await?;
let context = EngineContext {};
let open_request = OpenTableRequest {
catalog_name: t.catalog_name.clone(),
schema_name: t.schema_name.clone(),
table_name: t.table_name.clone(),
table_id: t.table_id,
region_numbers: vec![0],
};
let engine = self
.engine_manager
.engine(&t.engine)
.context(TableEngineNotFoundSnafu {
engine_name: &t.engine,
})?;
let table_ref = engine
.open_table(&context, open_request)
.await
.with_context(|_| OpenTableSnafu {
table_info: format!(
"{}.{}.{}, id: {}",
&t.catalog_name, &t.schema_name, &t.table_name, t.table_id
),
})?
.with_context(|| TableNotFoundSnafu {
table_info: format!(
"{}.{}.{}, id: {}",
&t.catalog_name, &t.schema_name, &t.table_name, t.table_id
),
})?;
let register_request = RegisterTableRequest {
catalog: t.catalog_name.clone(),
schema: t.schema_name.clone(),
table_name: t.table_name.clone(),
table_id: t.table_id,
table: table_ref,
};
let _ = self.catalogs.register_table(register_request).await?;
Ok(())
}
async fn check_state(&self) -> Result<()> {
let started = self.init_lock.lock().await;
ensure!(
*started,
IllegalManagerStateSnafu {
msg: "Catalog manager not started",
}
);
Ok(())
}
async fn check_catalog_schema_exist(
&self,
catalog_name: &str,
schema_name: &str,
) -> Result<()> {
if !self.catalogs.catalog_exist(catalog_name).await? {
return CatalogNotFoundSnafu { catalog_name }.fail()?;
}
if !self
.catalogs
.schema_exist(catalog_name, schema_name)
.await?
{
return SchemaNotFoundSnafu {
catalog: catalog_name,
schema: schema_name,
}
.fail()?;
}
Ok(())
}
}
#[async_trait::async_trait]
impl TableIdProvider for LocalCatalogManager {
async fn next_table_id(&self) -> table::Result<TableId> {
Ok(self.next_table_id.fetch_add(1, Ordering::Relaxed))
}
}
#[async_trait::async_trait]
impl CatalogManager for LocalCatalogManager {
/// Start [LocalCatalogManager] to load all information from system catalog table.
/// Make sure table engine is initialized before starting [MemoryCatalogManager].
async fn start(&self) -> Result<()> {
self.init().await
}
async fn register_table(&self, request: RegisterTableRequest) -> Result<bool> {
self.check_state().await?;
let catalog_name = request.catalog.clone();
let schema_name = request.schema.clone();
self.check_catalog_schema_exist(&catalog_name, &schema_name)
.await?;
{
let _lock = self.register_lock.lock().await;
if let Some(existing) = self
.catalogs
.table(&request.catalog, &request.schema, &request.table_name)
.await?
{
if existing.table_info().ident.table_id != request.table_id {
error!(
"Unexpected table register request: {:?}, existing: {:?}",
request,
existing.table_info()
);
return TableExistsSnafu {
table: format_full_table_name(
&catalog_name,
&schema_name,
&request.table_name,
),
}
.fail();
}
// Try to register table with same table id, just ignore.
Ok(false)
} else {
// table does not exist
let engine = request.table.table_info().meta.engine.to_string();
let table_name = request.table_name.clone();
let table_id = request.table_id;
let _ = self.catalogs.register_table(request).await?;
let _ = self
.system
.register_table(
catalog_name.clone(),
schema_name.clone(),
table_name,
table_id,
engine,
)
.await?;
increment_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_TABLE_COUNT,
1.0,
&[crate::metrics::db_label(&catalog_name, &schema_name)],
);
Ok(true)
}
}
}
async fn rename_table(&self, request: RenameTableRequest) -> Result<bool> {
self.check_state().await?;
let catalog_name = &request.catalog;
let schema_name = &request.schema;
self.check_catalog_schema_exist(catalog_name, schema_name)
.await?;
ensure!(
self.catalogs
.table(catalog_name, schema_name, &request.new_table_name)
.await?
.is_none(),
TableExistsSnafu {
table: &request.new_table_name
}
);
let _lock = self.register_lock.lock().await;
let old_table = self
.catalogs
.table(catalog_name, schema_name, &request.table_name)
.await?
.context(TableNotExistSnafu {
table: &request.table_name,
})?;
let engine = old_table.table_info().meta.engine.to_string();
// rename table in system catalog
let _ = self
.system
.register_table(
catalog_name.clone(),
schema_name.clone(),
request.new_table_name.clone(),
request.table_id,
engine,
)
.await?;
self.catalogs.rename_table(request).await
}
async fn deregister_table(&self, request: DeregisterTableRequest) -> Result<()> {
self.check_state().await?;
{
let _ = self.register_lock.lock().await;
let DeregisterTableRequest {
catalog,
schema,
table_name,
} = &request;
let table_id = self
.catalogs
.table(catalog, schema, table_name)
.await?
.with_context(|| error::TableNotExistSnafu {
table: format_full_table_name(catalog, schema, table_name),
})?
.table_info()
.ident
.table_id;
self.system.deregister_table(&request, table_id).await?;
self.catalogs.deregister_table(request).await
}
}
async fn register_schema(&self, request: RegisterSchemaRequest) -> Result<bool> {
self.check_state().await?;
let catalog_name = &request.catalog;
let schema_name = &request.schema;
if !self.catalogs.catalog_exist(catalog_name).await? {
return CatalogNotFoundSnafu { catalog_name }.fail()?;
}
{
let _lock = self.register_lock.lock().await;
ensure!(
!self
.catalogs
.schema_exist(catalog_name, schema_name)
.await?,
SchemaExistsSnafu {
schema: schema_name,
}
);
let _ = self
.system
.register_schema(request.catalog.clone(), schema_name.clone())
.await?;
self.catalogs.register_schema_sync(request)
}
}
async fn deregister_schema(&self, _request: DeregisterSchemaRequest) -> Result<bool> {
UnimplementedSnafu {
operation: "deregister schema",
}
.fail()
}
async fn register_system_table(&self, request: RegisterSystemTableRequest) -> Result<()> {
let catalog_name = request.create_table_request.catalog_name.clone();
let schema_name = request.create_table_request.schema_name.clone();
let mut sys_table_requests = self.system_table_requests.lock().await;
sys_table_requests.push(request);
increment_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_TABLE_COUNT,
1.0,
&[crate::metrics::db_label(&catalog_name, &schema_name)],
);
Ok(())
}
async fn schema_exist(&self, catalog: &str, schema: &str) -> Result<bool> {
self.catalogs.schema_exist(catalog, schema).await
}
async fn table(
&self,
catalog_name: &str,
schema_name: &str,
table_name: &str,
) -> Result<Option<TableRef>> {
self.catalogs
.table(catalog_name, schema_name, table_name)
.await
}
async fn catalog_exist(&self, catalog: &str) -> Result<bool> {
if catalog.eq_ignore_ascii_case(SYSTEM_CATALOG_NAME) {
Ok(true)
} else {
self.catalogs.catalog_exist(catalog).await
}
}
async fn table_exist(&self, catalog: &str, schema: &str, table: &str) -> Result<bool> {
self.catalogs.table_exist(catalog, schema, table).await
}
async fn catalog_names(&self) -> Result<Vec<String>> {
self.catalogs.catalog_names().await
}
async fn schema_names(&self, catalog_name: &str) -> Result<Vec<String>> {
self.catalogs.schema_names(catalog_name).await
}
async fn table_names(&self, catalog_name: &str, schema_name: &str) -> Result<Vec<String>> {
self.catalogs.table_names(catalog_name, schema_name).await
}
async fn register_catalog(self: Arc<Self>, name: String) -> Result<bool> {
self.catalogs.clone().register_catalog(name).await
}
fn as_any(&self) -> &dyn Any {
self
}
}
#[cfg(test)]
mod tests {
use std::assert_matches::assert_matches;
use mito::engine::MITO_ENGINE;
use super::*;
use crate::system::{CatalogEntry, SchemaEntry};
#[test]
fn test_sort_entry() {
let vec = vec![
Entry::Table(TableEntry {
catalog_name: "C1".to_string(),
schema_name: "S1".to_string(),
table_name: "T1".to_string(),
table_id: 1,
engine: MITO_ENGINE.to_string(),
is_deleted: false,
}),
Entry::Catalog(CatalogEntry {
catalog_name: "C2".to_string(),
}),
Entry::Schema(SchemaEntry {
catalog_name: "C1".to_string(),
schema_name: "S1".to_string(),
}),
Entry::Schema(SchemaEntry {
catalog_name: "C2".to_string(),
schema_name: "S2".to_string(),
}),
Entry::Catalog(CatalogEntry {
catalog_name: "".to_string(),
}),
Entry::Table(TableEntry {
catalog_name: "C1".to_string(),
schema_name: "S1".to_string(),
table_name: "T2".to_string(),
table_id: 2,
engine: MITO_ENGINE.to_string(),
is_deleted: false,
}),
];
let res = LocalCatalogManager::sort_entries(vec);
assert_matches!(res[0], Entry::Catalog(..));
assert_matches!(res[1], Entry::Catalog(..));
assert_matches!(res[2], Entry::Schema(..));
assert_matches!(res[3], Entry::Schema(..));
assert_matches!(res[4], Entry::Table(..));
assert_matches!(res[5], Entry::Table(..));
}
}

View File

@@ -1,606 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::any::Any;
use std::collections::hash_map::Entry;
use std::collections::HashMap;
use std::sync::atomic::{AtomicU32, Ordering};
use std::sync::{Arc, RwLock, Weak};
use common_catalog::consts::{
DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME, MIN_USER_TABLE_ID,
};
use metrics::{decrement_gauge, increment_gauge};
use snafu::OptionExt;
use table::metadata::TableId;
use table::table::TableIdProvider;
use table::TableRef;
use crate::error::{
CatalogNotFoundSnafu, Result, SchemaNotFoundSnafu, TableExistsSnafu, TableNotFoundSnafu,
};
use crate::information_schema::InformationSchemaProvider;
use crate::{
CatalogManager, DeregisterSchemaRequest, DeregisterTableRequest, RegisterSchemaRequest,
RegisterSystemTableRequest, RegisterTableRequest, RenameTableRequest,
};
type SchemaEntries = HashMap<String, HashMap<String, TableRef>>;
/// Simple in-memory list of catalogs
pub struct MemoryCatalogManager {
/// Collection of catalogs containing schemas and ultimately Tables
pub catalogs: RwLock<HashMap<String, SchemaEntries>>,
pub table_id: AtomicU32,
}
#[async_trait::async_trait]
impl TableIdProvider for MemoryCatalogManager {
async fn next_table_id(&self) -> table::error::Result<TableId> {
Ok(self.table_id.fetch_add(1, Ordering::Relaxed))
}
}
#[async_trait::async_trait]
impl CatalogManager for MemoryCatalogManager {
async fn start(&self) -> Result<()> {
self.table_id.store(MIN_USER_TABLE_ID, Ordering::Relaxed);
Ok(())
}
async fn register_table(&self, request: RegisterTableRequest) -> Result<bool> {
self.register_table_sync(request)
}
async fn rename_table(&self, request: RenameTableRequest) -> Result<bool> {
let mut catalogs = self.catalogs.write().unwrap();
let schema = catalogs
.get_mut(&request.catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: &request.catalog,
})?
.get_mut(&request.schema)
.with_context(|| SchemaNotFoundSnafu {
catalog: &request.catalog,
schema: &request.schema,
})?;
// check old and new table names
if !schema.contains_key(&request.table_name) {
return TableNotFoundSnafu {
table_info: request.table_name.to_string(),
}
.fail()?;
}
if schema.contains_key(&request.new_table_name) {
return TableExistsSnafu {
table: &request.new_table_name,
}
.fail();
}
let table = schema.remove(&request.table_name).unwrap();
let _ = schema.insert(request.new_table_name, table);
Ok(true)
}
async fn deregister_table(&self, request: DeregisterTableRequest) -> Result<()> {
let mut catalogs = self.catalogs.write().unwrap();
let schema = catalogs
.get_mut(&request.catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: &request.catalog,
})?
.get_mut(&request.schema)
.with_context(|| SchemaNotFoundSnafu {
catalog: &request.catalog,
schema: &request.schema,
})?;
let result = schema.remove(&request.table_name);
if result.is_some() {
decrement_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_TABLE_COUNT,
1.0,
&[crate::metrics::db_label(&request.catalog, &request.schema)],
);
}
Ok(())
}
async fn register_schema(&self, request: RegisterSchemaRequest) -> Result<bool> {
self.register_schema_sync(request)
}
async fn deregister_schema(&self, request: DeregisterSchemaRequest) -> Result<bool> {
let mut catalogs = self.catalogs.write().unwrap();
let schemas = catalogs
.get_mut(&request.catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: &request.catalog,
})?;
let table_count = schemas
.remove(&request.schema)
.with_context(|| SchemaNotFoundSnafu {
catalog: &request.catalog,
schema: &request.schema,
})?
.len();
decrement_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_TABLE_COUNT,
table_count as f64,
&[crate::metrics::db_label(&request.catalog, &request.schema)],
);
decrement_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_SCHEMA_COUNT,
1.0,
&[crate::metrics::db_label(&request.catalog, &request.schema)],
);
Ok(true)
}
async fn register_system_table(&self, _request: RegisterSystemTableRequest) -> Result<()> {
// TODO(ruihang): support register system table request
Ok(())
}
async fn schema_exist(&self, catalog: &str, schema: &str) -> Result<bool> {
Ok(self
.catalogs
.read()
.unwrap()
.get(catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: catalog,
})?
.contains_key(schema))
}
async fn table(
&self,
catalog: &str,
schema: &str,
table_name: &str,
) -> Result<Option<TableRef>> {
let result = try {
self.catalogs
.read()
.unwrap()
.get(catalog)?
.get(schema)?
.get(table_name)
.cloned()?
};
Ok(result)
}
async fn catalog_exist(&self, catalog: &str) -> Result<bool> {
Ok(self.catalogs.read().unwrap().get(catalog).is_some())
}
async fn table_exist(&self, catalog: &str, schema: &str, table: &str) -> Result<bool> {
let catalogs = self.catalogs.read().unwrap();
Ok(catalogs
.get(catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: catalog,
})?
.get(schema)
.with_context(|| SchemaNotFoundSnafu { catalog, schema })?
.contains_key(table))
}
async fn catalog_names(&self) -> Result<Vec<String>> {
Ok(self.catalogs.read().unwrap().keys().cloned().collect())
}
async fn schema_names(&self, catalog_name: &str) -> Result<Vec<String>> {
Ok(self
.catalogs
.read()
.unwrap()
.get(catalog_name)
.with_context(|| CatalogNotFoundSnafu { catalog_name })?
.keys()
.cloned()
.collect())
}
async fn table_names(&self, catalog_name: &str, schema_name: &str) -> Result<Vec<String>> {
Ok(self
.catalogs
.read()
.unwrap()
.get(catalog_name)
.with_context(|| CatalogNotFoundSnafu { catalog_name })?
.get(schema_name)
.with_context(|| SchemaNotFoundSnafu {
catalog: catalog_name,
schema: schema_name,
})?
.keys()
.cloned()
.collect())
}
async fn register_catalog(self: Arc<Self>, name: String) -> Result<bool> {
self.register_catalog_sync(name)
}
fn as_any(&self) -> &dyn Any {
self
}
}
impl MemoryCatalogManager {
/// Create a manager with some default setups
/// (e.g. default catalog/schema and information schema)
pub fn with_default_setup() -> Arc<Self> {
let manager = Arc::new(Self {
table_id: AtomicU32::new(MIN_USER_TABLE_ID),
catalogs: Default::default(),
});
// Safety: default catalog/schema is registered in order so no CatalogNotFound error will occur
manager
.register_catalog_sync(DEFAULT_CATALOG_NAME.to_string())
.unwrap();
manager
.register_schema_sync(RegisterSchemaRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
})
.unwrap();
manager
}
/// Registers a catalog and return the catalog already exist
pub fn register_catalog_if_absent(&self, name: String) -> bool {
let mut catalogs = self.catalogs.write().unwrap();
let entry = catalogs.entry(name);
match entry {
Entry::Occupied(_) => true,
Entry::Vacant(v) => {
let _ = v.insert(HashMap::new());
false
}
}
}
pub fn register_catalog_sync(self: &Arc<Self>, name: String) -> Result<bool> {
let mut catalogs = self.catalogs.write().unwrap();
match catalogs.entry(name.clone()) {
Entry::Vacant(e) => {
let catalog = self.create_catalog_entry(name);
e.insert(catalog);
increment_gauge!(crate::metrics::METRIC_CATALOG_MANAGER_CATALOG_COUNT, 1.0);
Ok(true)
}
Entry::Occupied(_) => Ok(false),
}
}
pub fn register_schema_sync(&self, request: RegisterSchemaRequest) -> Result<bool> {
let mut catalogs = self.catalogs.write().unwrap();
let catalog = catalogs
.get_mut(&request.catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: &request.catalog,
})?;
match catalog.entry(request.schema) {
Entry::Vacant(e) => {
e.insert(HashMap::new());
increment_gauge!(crate::metrics::METRIC_CATALOG_MANAGER_SCHEMA_COUNT, 1.0);
Ok(true)
}
Entry::Occupied(_) => Ok(false),
}
}
pub fn register_table_sync(&self, request: RegisterTableRequest) -> Result<bool> {
let mut catalogs = self.catalogs.write().unwrap();
let schema = catalogs
.get_mut(&request.catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: &request.catalog,
})?
.get_mut(&request.schema)
.with_context(|| SchemaNotFoundSnafu {
catalog: &request.catalog,
schema: &request.schema,
})?;
if schema.contains_key(&request.table_name) {
return TableExistsSnafu {
table: &request.table_name,
}
.fail();
}
schema.insert(request.table_name, request.table);
increment_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_TABLE_COUNT,
1.0,
&[crate::metrics::db_label(&request.catalog, &request.schema)],
);
Ok(true)
}
fn create_catalog_entry(self: &Arc<Self>, catalog: String) -> SchemaEntries {
let information_schema = InformationSchemaProvider::build(
catalog,
Arc::downgrade(self) as Weak<dyn CatalogManager>,
);
let mut catalog = HashMap::new();
catalog.insert(INFORMATION_SCHEMA_NAME.to_string(), information_schema);
catalog
}
#[cfg(any(test, feature = "testing"))]
pub fn new_with_table(table: TableRef) -> Arc<Self> {
let manager = Self::with_default_setup();
let request = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table.table_info().name.clone(),
table_id: table.table_info().ident.table_id,
table,
};
let _ = manager.register_table_sync(request).unwrap();
manager
}
}
/// Create a memory catalog list contains a numbers table for test
pub fn new_memory_catalog_manager() -> Result<Arc<MemoryCatalogManager>> {
Ok(MemoryCatalogManager::with_default_setup())
}
#[cfg(test)]
mod tests {
use common_catalog::consts::*;
use common_error::ext::ErrorExt;
use common_error::status_code::StatusCode;
use table::table::numbers::{NumbersTable, NUMBERS_TABLE_NAME};
use super::*;
#[tokio::test]
async fn test_new_memory_catalog_list() {
let catalog_list = new_memory_catalog_manager().unwrap();
let register_request = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: NUMBERS_TABLE_NAME.to_string(),
table_id: NUMBERS_TABLE_ID,
table: Arc::new(NumbersTable::default()),
};
let _ = catalog_list.register_table(register_request).await.unwrap();
let table = catalog_list
.table(
DEFAULT_CATALOG_NAME,
DEFAULT_SCHEMA_NAME,
NUMBERS_TABLE_NAME,
)
.await
.unwrap();
let _ = table.unwrap();
assert!(catalog_list
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, "not_exists")
.await
.unwrap()
.is_none());
}
#[tokio::test]
async fn test_mem_manager_rename_table() {
let catalog = MemoryCatalogManager::with_default_setup();
let table_name = "test_table";
assert!(!catalog
.table_exist(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap());
// register test table
let table_id = 2333;
let register_request = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
table_id,
table: Arc::new(NumbersTable::new(table_id)),
};
assert!(catalog.register_table(register_request).await.unwrap());
assert!(catalog
.table_exist(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap());
// rename test table
let new_table_name = "test_table_renamed";
let rename_request = RenameTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
new_table_name: new_table_name.to_string(),
table_id,
};
let _ = catalog.rename_table(rename_request).await.unwrap();
// test old table name not exist
assert!(!catalog
.table_exist(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap());
// test new table name exists
assert!(catalog
.table_exist(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, new_table_name)
.await
.unwrap());
let registered_table = catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, new_table_name)
.await
.unwrap()
.unwrap();
assert_eq!(registered_table.table_info().ident.table_id, table_id);
let dup_register_request = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: new_table_name.to_string(),
table_id: table_id + 1,
table: Arc::new(NumbersTable::new(table_id + 1)),
};
let result = catalog.register_table(dup_register_request).await;
let err = result.err().unwrap();
assert_eq!(StatusCode::TableAlreadyExists, err.status_code());
}
#[tokio::test]
async fn test_catalog_rename_table() {
let catalog = MemoryCatalogManager::with_default_setup();
let table_name = "num";
let table_id = 2333;
let table: TableRef = Arc::new(NumbersTable::new(table_id));
// register table
let register_table_req = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
table_id,
table,
};
assert!(catalog.register_table(register_table_req).await.unwrap());
assert!(catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap()
.is_some());
// rename table
let new_table_name = "numbers_new";
let rename_table_req = RenameTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
new_table_name: new_table_name.to_string(),
table_id,
};
assert!(catalog.rename_table(rename_table_req).await.unwrap());
assert!(catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap()
.is_none());
assert!(catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, new_table_name)
.await
.unwrap()
.is_some());
let registered_table = catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, new_table_name)
.await
.unwrap()
.unwrap();
assert_eq!(registered_table.table_info().ident.table_id, table_id);
}
#[test]
pub fn test_register_if_absent() {
let list = MemoryCatalogManager::with_default_setup();
assert!(!list.register_catalog_if_absent("test_catalog".to_string(),));
assert!(list.register_catalog_if_absent("test_catalog".to_string()));
}
#[tokio::test]
pub async fn test_catalog_deregister_table() {
let catalog = MemoryCatalogManager::with_default_setup();
let table_name = "foo_table";
let register_table_req = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
table_id: 2333,
table: Arc::new(NumbersTable::default()),
};
let _ = catalog.register_table(register_table_req).await.unwrap();
assert!(catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap()
.is_some());
let deregister_table_req = DeregisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
};
catalog
.deregister_table(deregister_table_req)
.await
.unwrap();
assert!(catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap()
.is_none());
}
#[tokio::test]
async fn test_catalog_deregister_schema() {
let catalog = MemoryCatalogManager::with_default_setup();
// Registers a catalog, a schema, and a table.
let catalog_name = "foo_catalog".to_string();
let schema_name = "foo_schema".to_string();
let table_name = "foo_table".to_string();
let schema = RegisterSchemaRequest {
catalog: catalog_name.clone(),
schema: schema_name.clone(),
};
let table = RegisterTableRequest {
catalog: catalog_name.clone(),
schema: schema_name.clone(),
table_name,
table_id: 0,
table: Arc::new(NumbersTable::default()),
};
catalog
.clone()
.register_catalog(catalog_name.clone())
.await
.unwrap();
catalog.register_schema(schema).await.unwrap();
catalog.register_table(table).await.unwrap();
let request = DeregisterSchemaRequest {
catalog: catalog_name.clone(),
schema: schema_name.clone(),
};
assert!(catalog.deregister_schema(request).await.unwrap());
assert!(!catalog
.schema_exist(&catalog_name, &schema_name)
.await
.unwrap());
}
}

View File

@@ -12,9 +12,6 @@
// See the License for the specific language governing permissions and
// limitations under the License.
pub mod config;
pub mod engine;
pub mod error;
mod manifest;
mod metrics;
pub mod table;
pub mod manager;
pub use manager::{new_memory_catalog_manager, MemoryCatalogManager};

View File

@@ -0,0 +1,369 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::any::Any;
use std::collections::hash_map::Entry;
use std::collections::HashMap;
use std::sync::{Arc, RwLock, Weak};
use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME};
use metrics::{decrement_gauge, increment_gauge};
use snafu::OptionExt;
use table::TableRef;
use crate::error::{CatalogNotFoundSnafu, Result, SchemaNotFoundSnafu, TableExistsSnafu};
use crate::information_schema::InformationSchemaProvider;
use crate::{CatalogManager, DeregisterTableRequest, RegisterSchemaRequest, RegisterTableRequest};
type SchemaEntries = HashMap<String, HashMap<String, TableRef>>;
/// Simple in-memory list of catalogs
#[derive(Clone)]
pub struct MemoryCatalogManager {
/// Collection of catalogs containing schemas and ultimately Tables
catalogs: Arc<RwLock<HashMap<String, SchemaEntries>>>,
}
#[async_trait::async_trait]
impl CatalogManager for MemoryCatalogManager {
async fn schema_exists(&self, catalog: &str, schema: &str) -> Result<bool> {
self.schema_exist_sync(catalog, schema)
}
async fn table(
&self,
catalog: &str,
schema: &str,
table_name: &str,
) -> Result<Option<TableRef>> {
let result = try {
self.catalogs
.read()
.unwrap()
.get(catalog)?
.get(schema)?
.get(table_name)
.cloned()?
};
Ok(result)
}
async fn catalog_exists(&self, catalog: &str) -> Result<bool> {
self.catalog_exist_sync(catalog)
}
async fn table_exists(&self, catalog: &str, schema: &str, table: &str) -> Result<bool> {
let catalogs = self.catalogs.read().unwrap();
Ok(catalogs
.get(catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: catalog,
})?
.get(schema)
.with_context(|| SchemaNotFoundSnafu { catalog, schema })?
.contains_key(table))
}
async fn catalog_names(&self) -> Result<Vec<String>> {
Ok(self.catalogs.read().unwrap().keys().cloned().collect())
}
async fn schema_names(&self, catalog_name: &str) -> Result<Vec<String>> {
Ok(self
.catalogs
.read()
.unwrap()
.get(catalog_name)
.with_context(|| CatalogNotFoundSnafu { catalog_name })?
.keys()
.cloned()
.collect())
}
async fn table_names(&self, catalog_name: &str, schema_name: &str) -> Result<Vec<String>> {
Ok(self
.catalogs
.read()
.unwrap()
.get(catalog_name)
.with_context(|| CatalogNotFoundSnafu { catalog_name })?
.get(schema_name)
.with_context(|| SchemaNotFoundSnafu {
catalog: catalog_name,
schema: schema_name,
})?
.keys()
.cloned()
.collect())
}
fn as_any(&self) -> &dyn Any {
self
}
}
impl MemoryCatalogManager {
pub fn new() -> Arc<Self> {
Arc::new(Self {
catalogs: Default::default(),
})
}
/// Creates a manager with some default setups
/// (e.g. default catalog/schema and information schema)
pub fn with_default_setup() -> Arc<Self> {
let manager = Arc::new(Self {
catalogs: Default::default(),
});
// Safety: default catalog/schema is registered in order so no CatalogNotFound error will occur
manager.register_catalog_sync(DEFAULT_CATALOG_NAME).unwrap();
manager
.register_schema_sync(RegisterSchemaRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
})
.unwrap();
manager
}
fn schema_exist_sync(&self, catalog: &str, schema: &str) -> Result<bool> {
Ok(self
.catalogs
.read()
.unwrap()
.get(catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: catalog,
})?
.contains_key(schema))
}
fn catalog_exist_sync(&self, catalog: &str) -> Result<bool> {
Ok(self.catalogs.read().unwrap().get(catalog).is_some())
}
/// Registers a catalog if it does not exist and returns false if the schema exists.
pub fn register_catalog_sync(&self, name: &str) -> Result<bool> {
let name = name.to_string();
let mut catalogs = self.catalogs.write().unwrap();
match catalogs.entry(name.clone()) {
Entry::Vacant(e) => {
let arc_self = Arc::new(self.clone());
let catalog = arc_self.create_catalog_entry(name);
e.insert(catalog);
increment_gauge!(crate::metrics::METRIC_CATALOG_MANAGER_CATALOG_COUNT, 1.0);
Ok(true)
}
Entry::Occupied(_) => Ok(false),
}
}
pub fn deregister_table_sync(&self, request: DeregisterTableRequest) -> Result<()> {
let mut catalogs = self.catalogs.write().unwrap();
let schema = catalogs
.get_mut(&request.catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: &request.catalog,
})?
.get_mut(&request.schema)
.with_context(|| SchemaNotFoundSnafu {
catalog: &request.catalog,
schema: &request.schema,
})?;
let result = schema.remove(&request.table_name);
if result.is_some() {
decrement_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_TABLE_COUNT,
1.0,
&[crate::metrics::db_label(&request.catalog, &request.schema)],
);
}
Ok(())
}
/// Registers a schema if it does not exist.
/// It returns an error if the catalog does not exist,
/// and returns false if the schema exists.
pub fn register_schema_sync(&self, request: RegisterSchemaRequest) -> Result<bool> {
let mut catalogs = self.catalogs.write().unwrap();
let catalog = catalogs
.get_mut(&request.catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: &request.catalog,
})?;
match catalog.entry(request.schema) {
Entry::Vacant(e) => {
e.insert(HashMap::new());
increment_gauge!(crate::metrics::METRIC_CATALOG_MANAGER_SCHEMA_COUNT, 1.0);
Ok(true)
}
Entry::Occupied(_) => Ok(false),
}
}
/// Registers a schema and returns an error if the catalog or schema does not exist.
pub fn register_table_sync(&self, request: RegisterTableRequest) -> Result<bool> {
let mut catalogs = self.catalogs.write().unwrap();
let schema = catalogs
.get_mut(&request.catalog)
.with_context(|| CatalogNotFoundSnafu {
catalog_name: &request.catalog,
})?
.get_mut(&request.schema)
.with_context(|| SchemaNotFoundSnafu {
catalog: &request.catalog,
schema: &request.schema,
})?;
if schema.contains_key(&request.table_name) {
return TableExistsSnafu {
table: &request.table_name,
}
.fail();
}
schema.insert(request.table_name, request.table);
increment_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_TABLE_COUNT,
1.0,
&[crate::metrics::db_label(&request.catalog, &request.schema)],
);
Ok(true)
}
fn create_catalog_entry(self: &Arc<Self>, catalog: String) -> SchemaEntries {
let information_schema = InformationSchemaProvider::build(
catalog,
Arc::downgrade(self) as Weak<dyn CatalogManager>,
);
let mut catalog = HashMap::new();
catalog.insert(INFORMATION_SCHEMA_NAME.to_string(), information_schema);
catalog
}
#[cfg(any(test, feature = "testing"))]
pub fn new_with_table(table: TableRef) -> Arc<Self> {
let manager = Self::with_default_setup();
let catalog = &table.table_info().catalog_name;
let schema = &table.table_info().schema_name;
if !manager.catalog_exist_sync(catalog).unwrap() {
manager.register_catalog_sync(catalog).unwrap();
}
if !manager.schema_exist_sync(catalog, schema).unwrap() {
manager
.register_schema_sync(RegisterSchemaRequest {
catalog: catalog.to_string(),
schema: schema.to_string(),
})
.unwrap();
}
let request = RegisterTableRequest {
catalog: catalog.to_string(),
schema: schema.to_string(),
table_name: table.table_info().name.clone(),
table_id: table.table_info().ident.table_id,
table,
};
let _ = manager.register_table_sync(request).unwrap();
manager
}
}
/// Create a memory catalog list contains a numbers table for test
pub fn new_memory_catalog_manager() -> Result<Arc<MemoryCatalogManager>> {
Ok(MemoryCatalogManager::with_default_setup())
}
#[cfg(test)]
mod tests {
use common_catalog::consts::*;
use table::table::numbers::{NumbersTable, NUMBERS_TABLE_NAME};
use super::*;
#[tokio::test]
async fn test_new_memory_catalog_list() {
let catalog_list = new_memory_catalog_manager().unwrap();
let register_request = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: NUMBERS_TABLE_NAME.to_string(),
table_id: NUMBERS_TABLE_ID,
table: NumbersTable::table(NUMBERS_TABLE_ID),
};
catalog_list.register_table_sync(register_request).unwrap();
let table = catalog_list
.table(
DEFAULT_CATALOG_NAME,
DEFAULT_SCHEMA_NAME,
NUMBERS_TABLE_NAME,
)
.await
.unwrap();
let _ = table.unwrap();
assert!(catalog_list
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, "not_exists")
.await
.unwrap()
.is_none());
}
#[test]
pub fn test_register_catalog_sync() {
let list = MemoryCatalogManager::with_default_setup();
assert!(list.register_catalog_sync("test_catalog").unwrap());
assert!(!list.register_catalog_sync("test_catalog").unwrap());
}
#[tokio::test]
pub async fn test_catalog_deregister_table() {
let catalog = MemoryCatalogManager::with_default_setup();
let table_name = "foo_table";
let register_table_req = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
table_id: 2333,
table: NumbersTable::table(2333),
};
catalog.register_table_sync(register_table_req).unwrap();
assert!(catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap()
.is_some());
let deregister_table_req = DeregisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
};
catalog.deregister_table_sync(deregister_table_req).unwrap();
assert!(catalog
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, table_name)
.await
.unwrap()
.is_none());
}
}

View File

@@ -1,434 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::any::Any;
use std::sync::Arc;
use async_trait::async_trait;
use common_catalog::consts::MITO_ENGINE;
use common_meta::helper::{CatalogKey, SchemaKey};
use common_meta::ident::TableIdent;
use common_meta::key::datanode_table::DatanodeTableValue;
use common_meta::key::TableMetadataManagerRef;
use common_meta::kv_backend::KvBackendRef;
use common_telemetry::{error, info, warn};
use metrics::increment_gauge;
use snafu::{ensure, OptionExt, ResultExt};
use table::engine::manager::TableEngineManagerRef;
use table::engine::EngineContext;
use table::requests::OpenTableRequest;
use table::TableRef;
use tokio::sync::Mutex;
use crate::error::{
OpenTableSnafu, ParallelOpenTableSnafu, Result, TableEngineNotFoundSnafu, TableExistsSnafu,
TableMetadataManagerSnafu, TableNotFoundSnafu, UnimplementedSnafu,
};
use crate::local::MemoryCatalogManager;
use crate::remote::region_alive_keeper::RegionAliveKeepers;
use crate::{
handle_system_table_request, CatalogManager, DeregisterSchemaRequest, DeregisterTableRequest,
RegisterSchemaRequest, RegisterSystemTableRequest, RegisterTableRequest, RenameTableRequest,
};
/// Catalog manager based on metasrv.
pub struct RemoteCatalogManager {
node_id: u64,
backend: KvBackendRef,
engine_manager: TableEngineManagerRef,
system_table_requests: Mutex<Vec<RegisterSystemTableRequest>>,
region_alive_keepers: Arc<RegionAliveKeepers>,
memory_catalog_manager: Arc<MemoryCatalogManager>,
table_metadata_manager: TableMetadataManagerRef,
}
impl RemoteCatalogManager {
pub fn new(
engine_manager: TableEngineManagerRef,
node_id: u64,
backend: KvBackendRef,
region_alive_keepers: Arc<RegionAliveKeepers>,
table_metadata_manager: TableMetadataManagerRef,
) -> Self {
Self {
engine_manager,
node_id,
backend,
system_table_requests: Default::default(),
region_alive_keepers,
memory_catalog_manager: MemoryCatalogManager::with_default_setup(),
table_metadata_manager,
}
}
async fn initiate_catalogs(&self) -> Result<()> {
let tables = self
.table_metadata_manager
.datanode_table_manager()
.tables(self.node_id)
.await
.context(TableMetadataManagerSnafu)?;
let joins = tables
.into_iter()
.map(|datanode_table_value| {
let engine_manager = self.engine_manager.clone();
let memory_catalog_manager = self.memory_catalog_manager.clone();
let table_metadata_manager = self.table_metadata_manager.clone();
common_runtime::spawn_bg(async move {
let table_id = datanode_table_value.table_id;
if let Err(e) = open_and_register_table(
engine_manager,
datanode_table_value,
memory_catalog_manager,
table_metadata_manager,
)
.await
{
// Note that we don't return error here if table opened failed. This is because
// we don't want those broken tables to impede the startup of Datanode.
// However, this could be changed in the future.
error!(e; "Failed to open or register table, id = {table_id}")
}
})
})
.collect::<Vec<_>>();
let _ = futures::future::try_join_all(joins)
.await
.context(ParallelOpenTableSnafu)?;
Ok(())
}
fn build_schema_key(&self, catalog_name: String, schema_name: String) -> SchemaKey {
SchemaKey {
catalog_name,
schema_name,
}
}
}
async fn open_and_register_table(
engine_manager: TableEngineManagerRef,
datanode_table_value: DatanodeTableValue,
memory_catalog_manager: Arc<MemoryCatalogManager>,
table_metadata_manager: TableMetadataManagerRef,
) -> Result<()> {
let context = EngineContext {};
let table_id = datanode_table_value.table_id;
let region_numbers = datanode_table_value.regions;
let table_info_value = table_metadata_manager
.table_info_manager()
.get(table_id)
.await
.context(TableMetadataManagerSnafu)?
.context(TableNotFoundSnafu {
table_info: format!("table id: {table_id}"),
})?;
let table_info = &table_info_value.table_info;
let catalog_name = table_info.catalog_name.clone();
let schema_name = table_info.schema_name.clone();
let table_name = table_info.name.clone();
let request = OpenTableRequest {
catalog_name: catalog_name.clone(),
schema_name: schema_name.clone(),
table_name: table_name.clone(),
table_id,
region_numbers: region_numbers.clone(),
};
let engine =
engine_manager
.engine(&table_info.meta.engine)
.context(TableEngineNotFoundSnafu {
engine_name: &table_info.meta.engine,
})?;
let table_ident = TableIdent {
catalog: catalog_name,
schema: schema_name,
table: table_name,
table_id,
engine: table_info.meta.engine.clone(),
};
let table = engine
.open_table(&context, request)
.await
.with_context(|_| OpenTableSnafu {
table_info: table_ident.to_string(),
})?
.with_context(|| TableNotFoundSnafu {
table_info: table_ident.to_string(),
})?;
info!("Successfully opened table, {table_ident}");
if !memory_catalog_manager
.catalog_exist(&table_ident.catalog)
.await?
{
memory_catalog_manager.register_catalog_sync(table_ident.catalog.clone())?;
}
if !memory_catalog_manager
.schema_exist(&table_ident.catalog, &table_ident.schema)
.await?
{
memory_catalog_manager.register_schema_sync(RegisterSchemaRequest {
catalog: table_ident.catalog.clone(),
schema: table_ident.schema.clone(),
})?;
}
let request = RegisterTableRequest {
catalog: table_ident.catalog.clone(),
schema: table_ident.schema.clone(),
table_name: table_ident.table.clone(),
table_id,
table,
};
let registered = memory_catalog_manager.register_table_sync(request)?;
ensure!(
registered,
TableExistsSnafu {
table: table_ident.to_string(),
}
);
info!("Successfully registered table, {table_ident}");
Ok(())
}
#[async_trait]
impl CatalogManager for RemoteCatalogManager {
async fn start(&self) -> Result<()> {
self.initiate_catalogs().await?;
let mut system_table_requests = self.system_table_requests.lock().await;
let engine = self
.engine_manager
.engine(MITO_ENGINE)
.context(TableEngineNotFoundSnafu {
engine_name: MITO_ENGINE,
})?;
handle_system_table_request(self, engine, &mut system_table_requests).await?;
info!("All system table opened");
Ok(())
}
async fn register_table(&self, request: RegisterTableRequest) -> Result<bool> {
let table = request.table.clone();
let registered = self.memory_catalog_manager.register_table_sync(request)?;
if registered {
let table_info = table.table_info();
let table_ident = TableIdent {
catalog: table_info.catalog_name.clone(),
schema: table_info.schema_name.clone(),
table: table_info.name.clone(),
table_id: table_info.table_id(),
engine: table_info.meta.engine.clone(),
};
self.region_alive_keepers
.register_table(table_ident, table)
.await?;
}
Ok(registered)
}
async fn deregister_table(&self, request: DeregisterTableRequest) -> Result<()> {
let Some(table) = self
.memory_catalog_manager
.table(&request.catalog, &request.schema, &request.table_name)
.await?
else {
return Ok(());
};
let table_info = table.table_info();
let table_ident = TableIdent {
catalog: request.catalog.clone(),
schema: request.schema.clone(),
table: request.table_name.clone(),
table_id: table_info.ident.table_id,
engine: table_info.meta.engine.clone(),
};
if let Some(keeper) = self
.region_alive_keepers
.deregister_table(&table_ident)
.await
{
warn!(
"Table {} is deregistered from region alive keepers",
keeper.table_ident(),
);
}
self.memory_catalog_manager.deregister_table(request).await
}
async fn register_schema(&self, request: RegisterSchemaRequest) -> Result<bool> {
self.memory_catalog_manager.register_schema_sync(request)
}
async fn deregister_schema(&self, _request: DeregisterSchemaRequest) -> Result<bool> {
UnimplementedSnafu {
operation: "deregister schema",
}
.fail()
}
async fn rename_table(&self, request: RenameTableRequest) -> Result<bool> {
self.memory_catalog_manager.rename_table(request).await
}
async fn register_system_table(&self, request: RegisterSystemTableRequest) -> Result<()> {
let catalog_name = request.create_table_request.catalog_name.clone();
let schema_name = request.create_table_request.schema_name.clone();
let mut requests = self.system_table_requests.lock().await;
requests.push(request);
increment_gauge!(
crate::metrics::METRIC_CATALOG_MANAGER_TABLE_COUNT,
1.0,
&[crate::metrics::db_label(&catalog_name, &schema_name)],
);
Ok(())
}
async fn schema_exist(&self, catalog: &str, schema: &str) -> Result<bool> {
if !self.catalog_exist(catalog).await? {
return Ok(false);
}
if self
.memory_catalog_manager
.schema_exist(catalog, schema)
.await?
{
return Ok(true);
}
let key = self
.build_schema_key(catalog.to_string(), schema.to_string())
.to_string();
let remote_schema_exists = self
.backend
.get(key.as_bytes())
.await
.context(TableMetadataManagerSnafu)?
.is_some();
// Create schema locally if remote schema exists. Since local schema is managed by memory
// catalog manager, creating a local schema is relatively cheap (just a HashMap).
// Besides, if this method ("schema_exist) is called, it's very likely that someone wants to
// create a table in this schema. We should create the schema now.
if remote_schema_exists
&& self
.memory_catalog_manager
.register_schema(RegisterSchemaRequest {
catalog: catalog.to_string(),
schema: schema.to_string(),
})
.await?
{
info!("register schema '{catalog}/{schema}' on demand");
}
Ok(remote_schema_exists)
}
async fn table(
&self,
catalog_name: &str,
schema_name: &str,
table_name: &str,
) -> Result<Option<TableRef>> {
self.memory_catalog_manager
.table(catalog_name, schema_name, table_name)
.await
}
async fn catalog_exist(&self, catalog: &str) -> Result<bool> {
if self.memory_catalog_manager.catalog_exist(catalog).await? {
return Ok(true);
}
let key = CatalogKey {
catalog_name: catalog.to_string(),
};
let remote_catalog_exists = self
.backend
.get(key.to_string().as_bytes())
.await
.context(TableMetadataManagerSnafu)?
.is_some();
// Create catalog locally if remote catalog exists. Since local catalog is managed by memory
// catalog manager, creating a local catalog is relatively cheap (just a HashMap).
// Besides, if this method ("catalog_exist) is called, it's very likely that someone wants to
// create a table in this catalog. We should create the catalog now.
if remote_catalog_exists
&& self
.memory_catalog_manager
.clone()
.register_catalog(catalog.to_string())
.await?
{
info!("register catalog '{catalog}' on demand");
}
Ok(remote_catalog_exists)
}
async fn table_exist(&self, catalog: &str, schema: &str, table: &str) -> Result<bool> {
if !self.catalog_exist(catalog).await? {
return Ok(false);
}
if !self.schema_exist(catalog, schema).await? {
return Ok(false);
}
self.memory_catalog_manager
.table_exist(catalog, schema, table)
.await
}
async fn catalog_names(&self) -> Result<Vec<String>> {
self.memory_catalog_manager.catalog_names().await
}
async fn schema_names(&self, catalog_name: &str) -> Result<Vec<String>> {
self.memory_catalog_manager.schema_names(catalog_name).await
}
async fn table_names(&self, catalog_name: &str, schema_name: &str) -> Result<Vec<String>> {
self.memory_catalog_manager
.table_names(catalog_name, schema_name)
.await
}
async fn register_catalog(self: Arc<Self>, name: String) -> Result<bool> {
self.memory_catalog_manager.register_catalog_sync(name)
}
fn as_any(&self) -> &dyn Any {
self
}
}

View File

@@ -1,826 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::collections::HashMap;
use std::future::Future;
use std::sync::atomic::{AtomicBool, Ordering};
use std::sync::Arc;
use async_trait::async_trait;
use common_meta::error::InvalidProtoMsgSnafu;
use common_meta::heartbeat::handler::{
HandleControl, HeartbeatResponseHandler, HeartbeatResponseHandlerContext,
};
use common_meta::ident::TableIdent;
use common_meta::RegionIdent;
use common_telemetry::{debug, error, info, warn};
use snafu::{OptionExt, ResultExt};
use store_api::storage::RegionNumber;
use table::engine::manager::TableEngineManagerRef;
use table::engine::{CloseTableResult, EngineContext, TableEngineRef};
use table::requests::CloseTableRequest;
use table::TableRef;
use tokio::sync::{mpsc, oneshot, Mutex};
use tokio::task::JoinHandle;
use tokio::time::{Duration, Instant};
use crate::error::{Result, TableEngineNotFoundSnafu};
/// [RegionAliveKeepers] manages all [RegionAliveKeeper] in a scope of tables.
pub struct RegionAliveKeepers {
table_engine_manager: TableEngineManagerRef,
keepers: Arc<Mutex<HashMap<TableIdent, Arc<RegionAliveKeeper>>>>,
heartbeat_interval_millis: u64,
started: AtomicBool,
/// The epoch when [RegionAliveKeepers] is created. It's used to get a monotonically non-decreasing
/// elapsed time when submitting heartbeats to Metasrv (because [Instant] is monotonically
/// non-decreasing). The heartbeat request will carry the duration since this epoch, and the
/// duration acts like an "invariant point" for region's keep alive lease.
epoch: Instant,
}
impl RegionAliveKeepers {
pub fn new(
table_engine_manager: TableEngineManagerRef,
heartbeat_interval_millis: u64,
) -> Self {
Self {
table_engine_manager,
keepers: Arc::new(Mutex::new(HashMap::new())),
heartbeat_interval_millis,
started: AtomicBool::new(false),
epoch: Instant::now(),
}
}
pub async fn find_keeper(&self, table_ident: &TableIdent) -> Option<Arc<RegionAliveKeeper>> {
self.keepers.lock().await.get(table_ident).cloned()
}
pub async fn register_table(&self, table_ident: TableIdent, table: TableRef) -> Result<()> {
let keeper = self.find_keeper(&table_ident).await;
if keeper.is_some() {
return Ok(());
}
let table_engine = self
.table_engine_manager
.engine(&table_ident.engine)
.context(TableEngineNotFoundSnafu {
engine_name: &table_ident.engine,
})?;
let keeper = Arc::new(RegionAliveKeeper::new(
table_engine,
table_ident.clone(),
self.heartbeat_interval_millis,
));
for r in table.table_info().meta.region_numbers.iter() {
keeper.register_region(*r).await;
}
let mut keepers = self.keepers.lock().await;
let _ = keepers.insert(table_ident.clone(), keeper.clone());
if self.started.load(Ordering::Relaxed) {
keeper.start().await;
info!("RegionAliveKeeper for table {table_ident} is started!");
} else {
info!("RegionAliveKeeper for table {table_ident} is registered but not started yet!");
}
Ok(())
}
pub async fn deregister_table(
&self,
table_ident: &TableIdent,
) -> Option<Arc<RegionAliveKeeper>> {
self.keepers.lock().await.remove(table_ident).map(|x| {
info!("Deregister RegionAliveKeeper for table {table_ident}");
x
})
}
pub async fn register_region(&self, region_ident: &RegionIdent) {
let table_ident = &region_ident.table_ident;
let Some(keeper) = self.find_keeper(table_ident).await else {
// Alive keeper could be affected by lagging msg, just warn and ignore.
warn!("Alive keeper for region {region_ident} is not found!");
return;
};
keeper.register_region(region_ident.region_number).await
}
pub async fn deregister_region(&self, region_ident: &RegionIdent) {
let table_ident = &region_ident.table_ident;
let Some(keeper) = self.find_keeper(table_ident).await else {
// Alive keeper could be affected by lagging msg, just warn and ignore.
warn!("Alive keeper for region {region_ident} is not found!");
return;
};
let _ = keeper.deregister_region(region_ident.region_number).await;
}
pub async fn start(&self) {
let keepers = self.keepers.lock().await;
for keeper in keepers.values() {
keeper.start().await;
}
self.started.store(true, Ordering::Relaxed);
info!(
"RegionAliveKeepers for tables {:?} are started!",
keepers.keys().map(|x| x.to_string()).collect::<Vec<_>>(),
);
}
pub fn epoch(&self) -> Instant {
self.epoch
}
}
#[async_trait]
impl HeartbeatResponseHandler for RegionAliveKeepers {
fn is_acceptable(&self, ctx: &HeartbeatResponseHandlerContext) -> bool {
!ctx.response.region_leases.is_empty()
}
async fn handle(
&self,
ctx: &mut HeartbeatResponseHandlerContext,
) -> common_meta::error::Result<HandleControl> {
let leases = ctx.response.region_leases.drain(..).collect::<Vec<_>>();
for lease in leases {
let table_ident: TableIdent = match lease
.table_ident
.context(InvalidProtoMsgSnafu {
err_msg: "'table_ident' is missing in RegionLease",
})
.and_then(|x| x.try_into())
{
Ok(x) => x,
Err(e) => {
error!(e; "");
continue;
}
};
let Some(keeper) = self.keepers.lock().await.get(&table_ident).cloned() else {
// Alive keeper could be affected by lagging msg, just warn and ignore.
warn!("Alive keeper for table {table_ident} is not found!");
continue;
};
let start_instant = self.epoch + Duration::from_millis(lease.duration_since_epoch);
let deadline = start_instant + Duration::from_secs(lease.lease_seconds);
keeper.keep_lived(lease.regions, deadline).await;
}
Ok(HandleControl::Continue)
}
}
/// [RegionAliveKeeper] starts a countdown for each region in a table. When deadline is reached,
/// the region will be closed.
/// The deadline is controlled by Metasrv. It works like "lease" for regions: a Datanode submits its
/// opened regions to Metasrv, in heartbeats. If Metasrv decides some region could be resided in this
/// Datanode, it will "extend" the region's "lease", with a deadline for [RegionAliveKeeper] to
/// countdown.
pub struct RegionAliveKeeper {
table_engine: TableEngineRef,
table_ident: TableIdent,
countdown_task_handles: Arc<Mutex<HashMap<RegionNumber, Arc<CountdownTaskHandle>>>>,
heartbeat_interval_millis: u64,
started: AtomicBool,
}
impl RegionAliveKeeper {
fn new(
table_engine: TableEngineRef,
table_ident: TableIdent,
heartbeat_interval_millis: u64,
) -> Self {
Self {
table_engine,
table_ident,
countdown_task_handles: Arc::new(Mutex::new(HashMap::new())),
heartbeat_interval_millis,
started: AtomicBool::new(false),
}
}
async fn find_handle(&self, region: &RegionNumber) -> Option<Arc<CountdownTaskHandle>> {
self.countdown_task_handles
.lock()
.await
.get(region)
.cloned()
}
async fn register_region(&self, region: RegionNumber) {
if self.find_handle(&region).await.is_some() {
return;
}
let countdown_task_handles = Arc::downgrade(&self.countdown_task_handles);
let on_task_finished = async move {
if let Some(x) = countdown_task_handles.upgrade() {
let _ = x.lock().await.remove(&region);
} // Else the countdown task handles map could be dropped because the keeper is dropped.
};
let handle = Arc::new(CountdownTaskHandle::new(
self.table_engine.clone(),
self.table_ident.clone(),
region,
|| on_task_finished,
));
let mut handles = self.countdown_task_handles.lock().await;
let _ = handles.insert(region, handle.clone());
if self.started.load(Ordering::Relaxed) {
handle.start(self.heartbeat_interval_millis).await;
info!(
"Region alive countdown for region {region} in table {} is started!",
self.table_ident
);
} else {
info!(
"Region alive countdown for region {region} in table {} is registered but not started yet!",
self.table_ident
);
}
}
async fn deregister_region(&self, region: RegionNumber) -> Option<Arc<CountdownTaskHandle>> {
self.countdown_task_handles
.lock()
.await
.remove(&region)
.map(|x| {
info!(
"Deregister alive countdown for region {region} in table {}",
self.table_ident
);
x
})
}
async fn start(&self) {
let handles = self.countdown_task_handles.lock().await;
for handle in handles.values() {
handle.start(self.heartbeat_interval_millis).await;
}
self.started.store(true, Ordering::Relaxed);
info!(
"Region alive countdowns for regions {:?} in table {} are started!",
handles.keys().copied().collect::<Vec<_>>(),
self.table_ident
);
}
async fn keep_lived(&self, designated_regions: Vec<RegionNumber>, deadline: Instant) {
for region in designated_regions {
if let Some(handle) = self.find_handle(&region).await {
handle.reset_deadline(deadline).await;
}
// Else the region alive keeper might be triggered by lagging messages, we can safely ignore it.
}
}
pub async fn deadline(&self, region: RegionNumber) -> Option<Instant> {
let mut deadline = None;
if let Some(handle) = self.find_handle(&region).await {
let (s, r) = oneshot::channel();
if handle.tx.send(CountdownCommand::Deadline(s)).await.is_ok() {
deadline = r.await.ok()
}
}
deadline
}
pub fn table_ident(&self) -> &TableIdent {
&self.table_ident
}
}
#[derive(Debug)]
enum CountdownCommand {
Start(u64),
Reset(Instant),
Deadline(oneshot::Sender<Instant>),
}
struct CountdownTaskHandle {
tx: mpsc::Sender<CountdownCommand>,
handler: JoinHandle<()>,
table_ident: TableIdent,
region: RegionNumber,
}
impl CountdownTaskHandle {
/// Creates a new [CountdownTaskHandle] and starts the countdown task.
/// # Params
/// - `on_task_finished`: a callback to be invoked when the task is finished. Note that it will not
/// be invoked if the task is cancelled (by dropping the handle). This is because we want something
/// meaningful to be done when the task is finished, e.g. deregister the handle from the map.
/// While dropping the handle does not necessarily mean the task is finished.
fn new<Fut>(
table_engine: TableEngineRef,
table_ident: TableIdent,
region: RegionNumber,
on_task_finished: impl FnOnce() -> Fut + Send + 'static,
) -> Self
where
Fut: Future<Output = ()> + Send,
{
let (tx, rx) = mpsc::channel(1024);
let mut countdown_task = CountdownTask {
table_engine,
table_ident: table_ident.clone(),
region,
rx,
};
let handler = common_runtime::spawn_bg(async move {
countdown_task.run().await;
on_task_finished().await;
});
Self {
tx,
handler,
table_ident,
region,
}
}
async fn start(&self, heartbeat_interval_millis: u64) {
if let Err(e) = self
.tx
.send(CountdownCommand::Start(heartbeat_interval_millis))
.await
{
warn!(
"Failed to start region alive keeper countdown: {e}. \
Maybe the task is stopped due to region been closed."
);
}
}
async fn reset_deadline(&self, deadline: Instant) {
if let Err(e) = self.tx.send(CountdownCommand::Reset(deadline)).await {
warn!(
"Failed to reset region alive keeper deadline: {e}. \
Maybe the task is stopped due to region been closed."
);
}
}
}
impl Drop for CountdownTaskHandle {
fn drop(&mut self) {
debug!(
"Aborting region alive countdown task for region {} in table {}",
self.region, self.table_ident,
);
self.handler.abort();
}
}
struct CountdownTask {
table_engine: TableEngineRef,
table_ident: TableIdent,
region: RegionNumber,
rx: mpsc::Receiver<CountdownCommand>,
}
impl CountdownTask {
async fn run(&mut self) {
// 30 years. See `Instant::far_future`.
let far_future = Instant::now() + Duration::from_secs(86400 * 365 * 30);
// Make sure the alive countdown is not gonna happen before heartbeat task is started (the
// "start countdown" command will be sent from heartbeat task).
let countdown = tokio::time::sleep_until(far_future);
tokio::pin!(countdown);
let region = &self.region;
let table_ident = &self.table_ident;
loop {
tokio::select! {
command = self.rx.recv() => {
match command {
Some(CountdownCommand::Start(heartbeat_interval_millis)) => {
// Set first deadline in 4 heartbeats (roughly after 20 seconds from now if heartbeat
// interval is set to default 5 seconds), to make Datanode and Metasrv more tolerable to
// network or other jitters during startup.
let first_deadline = Instant::now() + Duration::from_millis(heartbeat_interval_millis) * 4;
countdown.set(tokio::time::sleep_until(first_deadline));
},
Some(CountdownCommand::Reset(deadline)) => {
if countdown.deadline() < deadline {
debug!(
"Reset deadline of region {region} of table {table_ident} to approximately {} seconds later",
(deadline - Instant::now()).as_secs_f32(),
);
countdown.set(tokio::time::sleep_until(deadline));
}
// Else the countdown could be either:
// - not started yet;
// - during startup protection;
// - received a lagging heartbeat message.
// All can be safely ignored.
},
None => {
info!(
"The handle of countdown task for region {region} of table {table_ident} \
is dropped, RegionAliveKeeper out."
);
break;
},
Some(CountdownCommand::Deadline(tx)) => {
let _ = tx.send(countdown.deadline());
}
}
}
() = &mut countdown => {
let result = self.close_region().await;
warn!(
"Region {region} of table {table_ident} is closed, result: {result:?}. \
RegionAliveKeeper out.",
);
break;
}
}
}
}
async fn close_region(&self) -> CloseTableResult {
let ctx = EngineContext::default();
let region = self.region;
let table_ident = &self.table_ident;
loop {
let request = CloseTableRequest {
catalog_name: table_ident.catalog.clone(),
schema_name: table_ident.schema.clone(),
table_name: table_ident.table.clone(),
table_id: table_ident.table_id,
region_numbers: vec![region],
flush: true,
};
match self.table_engine.close_table(&ctx, request).await {
Ok(result) => return result,
// If region is failed to close, immediately retry. Maybe we should panic instead?
Err(e) => error!(e;
"Failed to close region {region} of table {table_ident}. \
For the integrity of data, retry closing and retry without wait.",
),
}
}
}
}
#[cfg(test)]
mod test {
use std::sync::atomic::{AtomicBool, Ordering};
use std::sync::Arc;
use api::v1::meta::{HeartbeatResponse, RegionLease};
use common_meta::heartbeat::mailbox::HeartbeatMailbox;
use datatypes::schema::RawSchema;
use table::engine::manager::MemoryTableEngineManager;
use table::engine::TableEngine;
use table::requests::{CreateTableRequest, TableOptions};
use table::test_util::EmptyTable;
use super::*;
use crate::remote::mock::MockTableEngine;
async fn prepare_keepers() -> (TableIdent, RegionAliveKeepers) {
let table_engine = Arc::new(MockTableEngine::default());
let table_engine_manager = Arc::new(MemoryTableEngineManager::new(table_engine));
let keepers = RegionAliveKeepers::new(table_engine_manager, 5000);
let catalog = "my_catalog";
let schema = "my_schema";
let table = "my_table";
let table_ident = TableIdent {
catalog: catalog.to_string(),
schema: schema.to_string(),
table: table.to_string(),
table_id: 1,
engine: "MockTableEngine".to_string(),
};
let table = Arc::new(EmptyTable::new(CreateTableRequest {
id: 1,
catalog_name: catalog.to_string(),
schema_name: schema.to_string(),
table_name: table.to_string(),
desc: None,
schema: RawSchema {
column_schemas: vec![],
timestamp_index: None,
version: 0,
},
region_numbers: vec![1, 2, 3],
primary_key_indices: vec![],
create_if_not_exists: false,
table_options: TableOptions::default(),
engine: "MockTableEngine".to_string(),
}));
keepers
.register_table(table_ident.clone(), table)
.await
.unwrap();
assert!(keepers.keepers.lock().await.contains_key(&table_ident));
(table_ident, keepers)
}
#[tokio::test(flavor = "multi_thread")]
async fn test_handle_heartbeat_response() {
let (table_ident, keepers) = prepare_keepers().await;
keepers.start().await;
let startup_protection_until = Instant::now() + Duration::from_secs(21);
let duration_since_epoch = (Instant::now() - keepers.epoch).as_millis() as _;
let lease_seconds = 100;
let response = HeartbeatResponse {
region_leases: vec![RegionLease {
table_ident: Some(table_ident.clone().into()),
regions: vec![1, 3], // Not extending region 2's lease time.
duration_since_epoch,
lease_seconds,
}],
..Default::default()
};
let keep_alive_until = keepers.epoch
+ Duration::from_millis(duration_since_epoch)
+ Duration::from_secs(lease_seconds);
let (tx, _) = mpsc::channel(8);
let mailbox = Arc::new(HeartbeatMailbox::new(tx));
let mut ctx = HeartbeatResponseHandlerContext::new(mailbox, response);
assert!(keepers.handle(&mut ctx).await.unwrap() == HandleControl::Continue);
// sleep to wait for background task spawned in `handle`
tokio::time::sleep(Duration::from_secs(1)).await;
async fn test(
keeper: &Arc<RegionAliveKeeper>,
region_number: RegionNumber,
startup_protection_until: Instant,
keep_alive_until: Instant,
is_kept_live: bool,
) {
let deadline = keeper.deadline(region_number).await.unwrap();
if is_kept_live {
assert!(deadline > startup_protection_until && deadline == keep_alive_until);
} else {
assert!(deadline <= startup_protection_until);
}
}
let keeper = &keepers
.keepers
.lock()
.await
.get(&table_ident)
.cloned()
.unwrap();
// Test region 1 and 3 is kept lived. Their deadlines are updated to desired instant.
test(keeper, 1, startup_protection_until, keep_alive_until, true).await;
test(keeper, 3, startup_protection_until, keep_alive_until, true).await;
// Test region 2 is not kept lived. It's deadline is not updated: still during startup protection period.
test(keeper, 2, startup_protection_until, keep_alive_until, false).await;
}
#[tokio::test(flavor = "multi_thread")]
async fn test_region_alive_keepers() {
let (table_ident, keepers) = prepare_keepers().await;
keepers
.register_region(&RegionIdent {
cluster_id: 1,
datanode_id: 1,
table_ident: table_ident.clone(),
region_number: 4,
})
.await;
keepers.start().await;
for keeper in keepers.keepers.lock().await.values() {
let regions = {
let handles = keeper.countdown_task_handles.lock().await;
handles.keys().copied().collect::<Vec<_>>()
};
for region in regions {
// assert countdown tasks are started
let deadline = keeper.deadline(region).await.unwrap();
assert!(deadline <= Instant::now() + Duration::from_secs(20));
}
}
keepers
.deregister_region(&RegionIdent {
cluster_id: 1,
datanode_id: 1,
table_ident: table_ident.clone(),
region_number: 1,
})
.await;
let mut regions = keepers
.find_keeper(&table_ident)
.await
.unwrap()
.countdown_task_handles
.lock()
.await
.keys()
.copied()
.collect::<Vec<_>>();
regions.sort();
assert_eq!(regions, vec![2, 3, 4]);
let keeper = keepers.deregister_table(&table_ident).await.unwrap();
assert!(Arc::try_unwrap(keeper).is_ok(), "keeper is not dropped");
assert!(keepers.keepers.lock().await.is_empty());
}
#[tokio::test(flavor = "multi_thread")]
async fn test_region_alive_keeper() {
let table_engine = Arc::new(MockTableEngine::default());
let table_ident = TableIdent {
catalog: "my_catalog".to_string(),
schema: "my_schema".to_string(),
table: "my_table".to_string(),
table_id: 1024,
engine: "mito".to_string(),
};
let keeper = RegionAliveKeeper::new(table_engine, table_ident, 1000);
let region = 1;
assert!(keeper.find_handle(&region).await.is_none());
keeper.register_region(region).await;
let _ = keeper.find_handle(&region).await.unwrap();
let ten_seconds_later = || Instant::now() + Duration::from_secs(10);
keeper.keep_lived(vec![1, 2, 3], ten_seconds_later()).await;
assert!(keeper.find_handle(&2).await.is_none());
assert!(keeper.find_handle(&3).await.is_none());
let far_future = Instant::now() + Duration::from_secs(86400 * 365 * 29);
// assert if keeper is not started, keep_lived is of no use
assert!(keeper.deadline(region).await.unwrap() > far_future);
keeper.start().await;
keeper.keep_lived(vec![1, 2, 3], ten_seconds_later()).await;
// assert keep_lived works if keeper is started
assert!(keeper.deadline(region).await.unwrap() <= ten_seconds_later());
let handle = keeper.deregister_region(region).await.unwrap();
assert!(Arc::try_unwrap(handle).is_ok(), "handle is not dropped");
assert!(keeper.find_handle(&region).await.is_none());
}
#[tokio::test(flavor = "multi_thread")]
async fn test_countdown_task_handle() {
let table_engine = Arc::new(MockTableEngine::default());
let table_ident = TableIdent {
catalog: "my_catalog".to_string(),
schema: "my_schema".to_string(),
table: "my_table".to_string(),
table_id: 1024,
engine: "mito".to_string(),
};
let finished = Arc::new(AtomicBool::new(false));
let finished_clone = finished.clone();
let handle = CountdownTaskHandle::new(
table_engine.clone(),
table_ident.clone(),
1,
|| async move { finished_clone.store(true, Ordering::Relaxed) },
);
let tx = handle.tx.clone();
// assert countdown task is running
tx.send(CountdownCommand::Start(5000)).await.unwrap();
assert!(!finished.load(Ordering::Relaxed));
drop(handle);
tokio::time::sleep(Duration::from_secs(1)).await;
// assert countdown task is stopped
assert!(tx
.try_send(CountdownCommand::Reset(
Instant::now() + Duration::from_secs(10)
))
.is_err());
// assert `on_task_finished` is not called (because the task is aborted by the handle's drop)
assert!(!finished.load(Ordering::Relaxed));
let finished = Arc::new(AtomicBool::new(false));
let finished_clone = finished.clone();
let handle = CountdownTaskHandle::new(table_engine, table_ident, 1, || async move {
finished_clone.store(true, Ordering::Relaxed)
});
handle.tx.send(CountdownCommand::Start(100)).await.unwrap();
tokio::time::sleep(Duration::from_secs(1)).await;
// assert `on_task_finished` is called when task is finished normally
assert!(finished.load(Ordering::Relaxed));
}
#[tokio::test(flavor = "multi_thread")]
async fn test_countdown_task_run() {
let ctx = &EngineContext::default();
let catalog = "my_catalog";
let schema = "my_schema";
let table = "my_table";
let table_id = 1;
let request = CreateTableRequest {
id: table_id,
catalog_name: catalog.to_string(),
schema_name: schema.to_string(),
table_name: table.to_string(),
desc: None,
schema: RawSchema {
column_schemas: vec![],
timestamp_index: None,
version: 0,
},
region_numbers: vec![],
primary_key_indices: vec![],
create_if_not_exists: false,
table_options: TableOptions::default(),
engine: "mito".to_string(),
};
let table_engine = Arc::new(MockTableEngine::default());
let _ = table_engine.create_table(ctx, request).await.unwrap();
let table_ident = TableIdent {
catalog: catalog.to_string(),
schema: schema.to_string(),
table: table.to_string(),
table_id,
engine: "mito".to_string(),
};
let (tx, rx) = mpsc::channel(10);
let mut task = CountdownTask {
table_engine: table_engine.clone(),
table_ident,
region: 1,
rx,
};
let _handle = common_runtime::spawn_bg(async move {
task.run().await;
});
async fn deadline(tx: &mpsc::Sender<CountdownCommand>) -> Instant {
let (s, r) = oneshot::channel();
tx.send(CountdownCommand::Deadline(s)).await.unwrap();
r.await.unwrap()
}
// if countdown task is not started, its deadline is set to far future
assert!(deadline(&tx).await > Instant::now() + Duration::from_secs(86400 * 365 * 29));
// start countdown in 250ms * 4 = 1s
tx.send(CountdownCommand::Start(250)).await.unwrap();
// assert deadline is correctly set
assert!(deadline(&tx).await <= Instant::now() + Duration::from_secs(1));
// reset countdown in 1.5s
tx.send(CountdownCommand::Reset(
Instant::now() + Duration::from_millis(1500),
))
.await
.unwrap();
// assert the table is closed after deadline is reached
assert!(table_engine.table_exists(ctx, table_id));
// spare 500ms for the task to close the table
tokio::time::sleep(Duration::from_millis(2000)).await;
assert!(!table_engine.table_exists(ctx, table_id));
}
}

View File

@@ -1,616 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::any::Any;
use std::collections::HashMap;
use std::sync::Arc;
use common_catalog::consts::{
DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME, MITO_ENGINE,
SYSTEM_CATALOG_NAME, SYSTEM_CATALOG_TABLE_ID, SYSTEM_CATALOG_TABLE_NAME,
};
use common_recordbatch::SendableRecordBatchStream;
use common_telemetry::debug;
use common_time::util;
use datatypes::prelude::{ConcreteDataType, ScalarVector, VectorRef};
use datatypes::schema::{ColumnSchema, RawSchema, SchemaRef};
use datatypes::vectors::{BinaryVector, TimestampMillisecondVector, UInt8Vector};
use serde::{Deserialize, Serialize};
use snafu::{ensure, OptionExt, ResultExt};
use store_api::storage::ScanRequest;
use table::engine::{EngineContext, TableEngineRef};
use table::metadata::{TableId, TableInfoRef, TableType};
use table::requests::{
CreateTableRequest, DeleteRequest, InsertRequest, OpenTableRequest, TableOptions,
};
use table::{Result as TableResult, Table, TableRef};
use crate::error::{
self, CreateSystemCatalogSnafu, EmptyValueSnafu, Error, InvalidEntryTypeSnafu, InvalidKeySnafu,
OpenSystemCatalogSnafu, Result, ValueDeserializeSnafu,
};
use crate::DeregisterTableRequest;
pub const ENTRY_TYPE_INDEX: usize = 0;
pub const KEY_INDEX: usize = 1;
pub const VALUE_INDEX: usize = 3;
pub struct SystemCatalogTable(TableRef);
#[async_trait::async_trait]
impl Table for SystemCatalogTable {
fn as_any(&self) -> &dyn Any {
self
}
fn schema(&self) -> SchemaRef {
self.0.schema()
}
async fn scan_to_stream(&self, request: ScanRequest) -> TableResult<SendableRecordBatchStream> {
self.0.scan_to_stream(request).await
}
/// Insert values into table.
async fn insert(&self, request: InsertRequest) -> TableResult<usize> {
self.0.insert(request).await
}
fn table_info(&self) -> TableInfoRef {
self.0.table_info()
}
fn table_type(&self) -> TableType {
self.0.table_type()
}
async fn delete(&self, request: DeleteRequest) -> TableResult<usize> {
self.0.delete(request).await
}
fn statistics(&self) -> Option<table::stats::TableStatistics> {
self.0.statistics()
}
}
impl SystemCatalogTable {
pub async fn new(engine: TableEngineRef) -> Result<Self> {
let request = OpenTableRequest {
catalog_name: SYSTEM_CATALOG_NAME.to_string(),
schema_name: INFORMATION_SCHEMA_NAME.to_string(),
table_name: SYSTEM_CATALOG_TABLE_NAME.to_string(),
table_id: SYSTEM_CATALOG_TABLE_ID,
region_numbers: vec![0],
};
let schema = build_system_catalog_schema();
let ctx = EngineContext::default();
if let Some(table) = engine
.open_table(&ctx, request)
.await
.context(OpenSystemCatalogSnafu)?
{
Ok(Self(table))
} else {
// system catalog table is not yet created, try to create
let request = CreateTableRequest {
id: SYSTEM_CATALOG_TABLE_ID,
catalog_name: SYSTEM_CATALOG_NAME.to_string(),
schema_name: INFORMATION_SCHEMA_NAME.to_string(),
table_name: SYSTEM_CATALOG_TABLE_NAME.to_string(),
desc: Some("System catalog table".to_string()),
schema,
region_numbers: vec![0],
primary_key_indices: vec![ENTRY_TYPE_INDEX, KEY_INDEX],
create_if_not_exists: true,
table_options: TableOptions::default(),
engine: engine.name().to_string(),
};
let table = engine
.create_table(&ctx, request)
.await
.context(CreateSystemCatalogSnafu)?;
Ok(Self(table))
}
}
/// Create a stream of all entries inside system catalog table
pub async fn records(&self) -> Result<SendableRecordBatchStream> {
let full_projection = None;
let scan_req = ScanRequest {
sequence: None,
projection: full_projection,
filters: vec![],
output_ordering: None,
limit: None,
};
let stream = self
.scan_to_stream(scan_req)
.await
.context(error::SystemCatalogTableScanSnafu)?;
Ok(stream)
}
}
/// Build system catalog table schema.
/// A system catalog table consists of 6 columns, namely
/// - entry_type: type of entry in current row, can be any variant of [EntryType].
/// - key: a binary encoded key of entry, differs according to different entry type.
/// - timestamp: currently not used.
/// - value: JSON-encoded value of entry's metadata.
/// - gmt_created: create time of this metadata.
/// - gmt_modified: last updated time of this metadata.
fn build_system_catalog_schema() -> RawSchema {
let cols = vec![
ColumnSchema::new(
"entry_type".to_string(),
ConcreteDataType::uint8_datatype(),
false,
),
ColumnSchema::new(
"key".to_string(),
ConcreteDataType::binary_datatype(),
false,
),
ColumnSchema::new(
"timestamp".to_string(),
ConcreteDataType::timestamp_millisecond_datatype(),
false,
)
.with_time_index(true),
ColumnSchema::new(
"value".to_string(),
ConcreteDataType::binary_datatype(),
false,
),
ColumnSchema::new(
"gmt_created".to_string(),
ConcreteDataType::timestamp_millisecond_datatype(),
false,
),
ColumnSchema::new(
"gmt_modified".to_string(),
ConcreteDataType::timestamp_millisecond_datatype(),
false,
),
];
RawSchema::new(cols)
}
/// Formats key string for table entry in system catalog
#[inline]
pub fn format_table_entry_key(catalog: &str, schema: &str, table_id: TableId) -> String {
format!("{catalog}.{schema}.{table_id}")
}
pub fn build_table_insert_request(
catalog: String,
schema: String,
table_name: String,
table_id: TableId,
engine: String,
) -> InsertRequest {
let entry_key = format_table_entry_key(&catalog, &schema, table_id);
build_insert_request(
EntryType::Table,
entry_key.as_bytes(),
serde_json::to_string(&TableEntryValue {
table_name,
engine,
is_deleted: false,
})
.unwrap()
.as_bytes(),
)
}
pub(crate) fn build_table_deletion_request(
request: &DeregisterTableRequest,
table_id: TableId,
) -> InsertRequest {
let entry_key = format_table_entry_key(&request.catalog, &request.schema, table_id);
build_insert_request(
EntryType::Table,
entry_key.as_bytes(),
serde_json::to_string(&TableEntryValue {
table_name: "".to_string(),
engine: "".to_string(),
is_deleted: true,
})
.unwrap()
.as_bytes(),
)
}
fn build_primary_key_columns(entry_type: EntryType, key: &[u8]) -> HashMap<String, VectorRef> {
HashMap::from([
(
"entry_type".to_string(),
Arc::new(UInt8Vector::from_slice([entry_type as u8])) as VectorRef,
),
(
"key".to_string(),
Arc::new(BinaryVector::from_slice(&[key])) as VectorRef,
),
(
"timestamp".to_string(),
// Timestamp in key part is intentionally left to 0
Arc::new(TimestampMillisecondVector::from_slice([0])) as VectorRef,
),
])
}
pub fn build_schema_insert_request(catalog_name: String, schema_name: String) -> InsertRequest {
let full_schema_name = format!("{catalog_name}.{schema_name}");
build_insert_request(
EntryType::Schema,
full_schema_name.as_bytes(),
serde_json::to_string(&SchemaEntryValue {})
.unwrap()
.as_bytes(),
)
}
pub fn build_insert_request(entry_type: EntryType, key: &[u8], value: &[u8]) -> InsertRequest {
let primary_key_columns = build_primary_key_columns(entry_type, key);
let mut columns_values = HashMap::with_capacity(6);
columns_values.extend(primary_key_columns);
let _ = columns_values.insert(
"value".to_string(),
Arc::new(BinaryVector::from_slice(&[value])) as _,
);
let now = util::current_time_millis();
let _ = columns_values.insert(
"gmt_created".to_string(),
Arc::new(TimestampMillisecondVector::from_slice([now])) as _,
);
let _ = columns_values.insert(
"gmt_modified".to_string(),
Arc::new(TimestampMillisecondVector::from_slice([now])) as _,
);
InsertRequest {
catalog_name: DEFAULT_CATALOG_NAME.to_string(),
schema_name: DEFAULT_SCHEMA_NAME.to_string(),
table_name: SYSTEM_CATALOG_TABLE_NAME.to_string(),
columns_values,
region_number: 0, // system catalog table has only one region
}
}
pub fn decode_system_catalog(
entry_type: Option<u8>,
key: Option<&[u8]>,
value: Option<&[u8]>,
) -> Result<Entry> {
debug!(
"Decode system catalog entry: {:?}, {:?}, {:?}",
entry_type, key, value
);
let entry_type = entry_type.context(InvalidKeySnafu { key: None })?;
let key = String::from_utf8_lossy(key.context(InvalidKeySnafu { key: None })?);
match EntryType::try_from(entry_type)? {
EntryType::Catalog => {
// As for catalog entry, the key is a string with format: `<catalog_name>`
// and the value is current not used.
let catalog_name = key.to_string();
Ok(Entry::Catalog(CatalogEntry { catalog_name }))
}
EntryType::Schema => {
// As for schema entry, the key is a string with format: `<catalog_name>.<schema_name>`
// and the value is current not used.
let schema_parts = key.split('.').collect::<Vec<_>>();
ensure!(
schema_parts.len() == 2,
InvalidKeySnafu {
key: Some(key.to_string())
}
);
Ok(Entry::Schema(SchemaEntry {
catalog_name: schema_parts[0].to_string(),
schema_name: schema_parts[1].to_string(),
}))
}
EntryType::Table => {
// As for table entry, the key is a string with format: `<catalog_name>.<schema_name>.<table_id>`
// and the value is a JSON string with format: `{"table_name": <table_name>}`
let table_parts = key.split('.').collect::<Vec<_>>();
ensure!(
table_parts.len() >= 3,
InvalidKeySnafu {
key: Some(key.to_string())
}
);
let value = value.context(EmptyValueSnafu)?;
debug!("Table meta value: {}", String::from_utf8_lossy(value));
let table_meta: TableEntryValue =
serde_json::from_slice(value).context(ValueDeserializeSnafu)?;
let table_id = table_parts[2].parse::<TableId>().unwrap();
Ok(Entry::Table(TableEntry {
catalog_name: table_parts[0].to_string(),
schema_name: table_parts[1].to_string(),
table_name: table_meta.table_name,
table_id,
engine: table_meta.engine,
is_deleted: table_meta.is_deleted,
}))
}
}
}
#[derive(Debug, Copy, Clone, PartialEq, Eq)]
pub enum EntryType {
Catalog = 1,
Schema = 2,
Table = 3,
}
impl TryFrom<u8> for EntryType {
type Error = Error;
fn try_from(value: u8) -> std::result::Result<Self, Self::Error> {
match value {
b if b == Self::Catalog as u8 => Ok(Self::Catalog),
b if b == Self::Schema as u8 => Ok(Self::Schema),
b if b == Self::Table as u8 => Ok(Self::Table),
b => InvalidEntryTypeSnafu {
entry_type: Some(b),
}
.fail(),
}
}
}
#[derive(Debug, PartialEq, Eq, Ord, PartialOrd)]
pub enum Entry {
Catalog(CatalogEntry),
Schema(SchemaEntry),
Table(TableEntry),
}
#[derive(Debug, PartialEq, Eq, Ord, PartialOrd)]
pub struct CatalogEntry {
pub catalog_name: String,
}
#[derive(Debug, PartialEq, Eq, Ord, PartialOrd)]
pub struct SchemaEntry {
pub catalog_name: String,
pub schema_name: String,
}
#[derive(Debug, Serialize, Deserialize, PartialEq, Eq)]
pub struct SchemaEntryValue;
#[derive(Debug, PartialEq, Eq, Ord, PartialOrd)]
pub struct TableEntry {
pub catalog_name: String,
pub schema_name: String,
pub table_name: String,
pub table_id: TableId,
pub engine: String,
pub is_deleted: bool,
}
#[derive(Debug, Serialize, Deserialize, PartialEq, Eq)]
pub struct TableEntryValue {
pub table_name: String,
#[serde(default = "mito_engine")]
pub engine: String,
#[serde(default = "not_deleted")]
pub is_deleted: bool,
}
fn mito_engine() -> String {
MITO_ENGINE.to_string()
}
fn not_deleted() -> bool {
false
}
#[cfg(test)]
mod tests {
use common_recordbatch::RecordBatches;
use common_test_util::temp_dir::{create_temp_dir, TempDir};
use datatypes::value::Value;
use log_store::NoopLogStore;
use mito::config::EngineConfig;
use mito::engine::{MitoEngine, MITO_ENGINE};
use object_store::ObjectStore;
use storage::compaction::noop::NoopCompactionScheduler;
use storage::config::EngineConfig as StorageEngineConfig;
use storage::EngineImpl;
use table::metadata::TableType;
use table::metadata::TableType::Base;
use super::*;
#[test]
pub fn test_decode_catalog_entry() {
let entry = decode_system_catalog(
Some(EntryType::Catalog as u8),
Some("some_catalog".as_bytes()),
None,
)
.unwrap();
if let Entry::Catalog(e) = entry {
assert_eq!("some_catalog", e.catalog_name);
} else {
panic!("Unexpected type: {entry:?}");
}
}
#[test]
pub fn test_decode_schema_entry() {
let entry = decode_system_catalog(
Some(EntryType::Schema as u8),
Some("some_catalog.some_schema".as_bytes()),
None,
)
.unwrap();
if let Entry::Schema(e) = entry {
assert_eq!("some_catalog", e.catalog_name);
assert_eq!("some_schema", e.schema_name);
} else {
panic!("Unexpected type: {entry:?}");
}
}
#[test]
pub fn test_decode_table() {
let entry = decode_system_catalog(
Some(EntryType::Table as u8),
Some("some_catalog.some_schema.42".as_bytes()),
Some("{\"table_name\":\"some_table\"}".as_bytes()),
)
.unwrap();
if let Entry::Table(e) = entry {
assert_eq!("some_catalog", e.catalog_name);
assert_eq!("some_schema", e.schema_name);
assert_eq!("some_table", e.table_name);
assert_eq!(42, e.table_id);
} else {
panic!("Unexpected type: {entry:?}");
}
}
#[test]
pub fn test_decode_mismatch() {
assert!(decode_system_catalog(
Some(EntryType::Table as u8),
Some("some_catalog.some_schema.42".as_bytes()),
None,
)
.is_err());
}
#[test]
pub fn test_entry_type() {
assert_eq!(EntryType::Catalog, EntryType::try_from(1).unwrap());
assert_eq!(EntryType::Schema, EntryType::try_from(2).unwrap());
assert_eq!(EntryType::Table, EntryType::try_from(3).unwrap());
assert!(EntryType::try_from(4).is_err());
}
pub async fn prepare_table_engine() -> (TempDir, TableEngineRef) {
let dir = create_temp_dir("system-table-test");
let store_dir = dir.path().to_string_lossy();
let mut builder = object_store::services::Fs::default();
let _ = builder.root(&store_dir);
let object_store = ObjectStore::new(builder).unwrap().finish();
let noop_compaction_scheduler = Arc::new(NoopCompactionScheduler::default());
let table_engine = Arc::new(MitoEngine::new(
EngineConfig::default(),
EngineImpl::new(
StorageEngineConfig::default(),
Arc::new(NoopLogStore),
object_store.clone(),
noop_compaction_scheduler,
)
.unwrap(),
object_store,
));
(dir, table_engine)
}
#[tokio::test]
async fn test_system_table_type() {
let (_dir, table_engine) = prepare_table_engine().await;
let system_table = SystemCatalogTable::new(table_engine).await.unwrap();
assert_eq!(Base, system_table.table_type());
}
#[tokio::test]
async fn test_system_table_info() {
let (_dir, table_engine) = prepare_table_engine().await;
let system_table = SystemCatalogTable::new(table_engine).await.unwrap();
let info = system_table.table_info();
assert_eq!(TableType::Base, info.table_type);
assert_eq!(SYSTEM_CATALOG_TABLE_NAME, info.name);
assert_eq!(SYSTEM_CATALOG_TABLE_ID, info.ident.table_id);
assert_eq!(SYSTEM_CATALOG_NAME, info.catalog_name);
assert_eq!(INFORMATION_SCHEMA_NAME, info.schema_name);
}
#[tokio::test]
async fn test_system_catalog_table_records() {
let (_, table_engine) = prepare_table_engine().await;
let catalog_table = SystemCatalogTable::new(table_engine).await.unwrap();
let table_insertion = build_table_insert_request(
DEFAULT_CATALOG_NAME.to_string(),
DEFAULT_SCHEMA_NAME.to_string(),
"my_table".to_string(),
1,
MITO_ENGINE.to_string(),
);
let result = catalog_table.insert(table_insertion).await.unwrap();
assert_eq!(result, 1);
let records = catalog_table.records().await.unwrap();
let mut batches = RecordBatches::try_collect(records).await.unwrap().take();
assert_eq!(batches.len(), 1);
let batch = batches.remove(0);
assert_eq!(batch.num_rows(), 1);
let row = batch.rows().next().unwrap();
let Value::UInt8(entry_type) = row[0] else {
unreachable!()
};
let Value::Binary(key) = row[1].clone() else {
unreachable!()
};
let Value::Binary(value) = row[3].clone() else {
unreachable!()
};
let entry = decode_system_catalog(Some(entry_type), Some(&*key), Some(&*value)).unwrap();
let expected = Entry::Table(TableEntry {
catalog_name: DEFAULT_CATALOG_NAME.to_string(),
schema_name: DEFAULT_SCHEMA_NAME.to_string(),
table_name: "my_table".to_string(),
table_id: 1,
engine: MITO_ENGINE.to_string(),
is_deleted: false,
});
assert_eq!(entry, expected);
let table_deletion = build_table_deletion_request(
&DeregisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: "my_table".to_string(),
},
1,
);
let result = catalog_table.insert(table_deletion).await.unwrap();
assert_eq!(result, 1);
let records = catalog_table.records().await.unwrap();
let batches = RecordBatches::try_collect(records).await.unwrap().take();
assert_eq!(batches.len(), 1);
}
}

View File

@@ -123,7 +123,7 @@ mod tests {
use session::context::QueryContext;
use super::*;
use crate::local::MemoryCatalogManager;
use crate::memory::MemoryCatalogManager;
#[test]
fn test_validate_table_ref() {

View File

@@ -1,101 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// The `tables` table in system catalog keeps a record of all tables created by user.
use std::sync::Arc;
use common_telemetry::logging;
use snafu::ResultExt;
use table::metadata::TableId;
use table::Table;
use crate::error::{self, InsertCatalogRecordSnafu, Result as CatalogResult};
use crate::system::{
build_schema_insert_request, build_table_deletion_request, build_table_insert_request,
SystemCatalogTable,
};
use crate::DeregisterTableRequest;
pub struct InformationSchema {
pub system: Arc<SystemCatalogTable>,
}
pub struct SystemCatalog {
pub information_schema: Arc<InformationSchema>,
}
impl SystemCatalog {
pub(crate) fn new(system: SystemCatalogTable) -> Self {
let schema = InformationSchema {
system: Arc::new(system),
};
Self {
information_schema: Arc::new(schema),
}
}
pub async fn register_table(
&self,
catalog: String,
schema: String,
table_name: String,
table_id: TableId,
engine: String,
) -> crate::error::Result<usize> {
let request = build_table_insert_request(catalog, schema, table_name, table_id, engine);
self.information_schema
.system
.insert(request)
.await
.context(InsertCatalogRecordSnafu)
}
pub(crate) async fn deregister_table(
&self,
request: &DeregisterTableRequest,
table_id: TableId,
) -> CatalogResult<()> {
self.information_schema
.system
.insert(build_table_deletion_request(request, table_id))
.await
.map(|x| {
if x != 1 {
let table = common_catalog::format_full_table_name(
&request.catalog,
&request.schema,
&request.table_name
);
logging::warn!("Failed to delete table record from information_schema, unexpected returned result: {x}, table: {table}");
}
})
.with_context(|_| error::DeregisterTableSnafu {
request: request.clone(),
})
}
pub async fn register_schema(
&self,
catalog: String,
schema: String,
) -> crate::error::Result<usize> {
let request = build_schema_insert_request(catalog, schema);
self.information_schema
.system
.insert(request)
.await
.context(InsertCatalogRecordSnafu)
}
}

View File

@@ -1,176 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#[cfg(test)]
mod tests {
use std::sync::Arc;
use catalog::local::LocalCatalogManager;
use catalog::{CatalogManager, RegisterTableRequest, RenameTableRequest};
use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
use common_telemetry::{error, info};
use common_test_util::temp_dir::TempDir;
use mito::config::EngineConfig;
use table::engine::manager::MemoryTableEngineManager;
use table::table::numbers::NumbersTable;
use table::TableRef;
use tokio::sync::Mutex;
async fn create_local_catalog_manager(
) -> Result<(TempDir, LocalCatalogManager), catalog::error::Error> {
let (dir, object_store) =
mito::table::test_util::new_test_object_store("setup_mock_engine_and_table").await;
let mock_engine = Arc::new(mito::table::test_util::MockMitoEngine::new(
EngineConfig::default(),
mito::table::test_util::MockEngine::default(),
object_store,
));
let engine_manager = Arc::new(MemoryTableEngineManager::new(mock_engine.clone()));
let catalog_manager = LocalCatalogManager::try_new(engine_manager).await.unwrap();
catalog_manager.start().await?;
Ok((dir, catalog_manager))
}
#[tokio::test]
async fn test_rename_table() {
common_telemetry::init_default_ut_logging();
let (_dir, catalog_manager) = create_local_catalog_manager().await.unwrap();
// register table
let table_name = "test_table";
let table_id = 42;
let table = Arc::new(NumbersTable::new(table_id));
let request = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
table_id,
table: table.clone(),
};
assert!(catalog_manager.register_table(request).await.unwrap());
// rename table
let new_table_name = "table_t";
let rename_table_req = RenameTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: table_name.to_string(),
new_table_name: new_table_name.to_string(),
table_id,
};
assert!(catalog_manager
.rename_table(rename_table_req)
.await
.unwrap());
let registered_table = catalog_manager
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, new_table_name)
.await
.unwrap()
.unwrap();
assert_eq!(registered_table.table_info().ident.table_id, table_id);
}
#[tokio::test]
async fn test_duplicate_register() {
let (_dir, catalog_manager) = create_local_catalog_manager().await.unwrap();
let request = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: "test_table".to_string(),
table_id: 42,
table: Arc::new(NumbersTable::new(42)),
};
assert!(catalog_manager
.register_table(request.clone())
.await
.unwrap());
// register table with same table id will succeed with 0 as return val.
assert!(!catalog_manager.register_table(request).await.unwrap());
let err = catalog_manager
.register_table(RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: "test_table".to_string(),
table_id: 43,
table: Arc::new(NumbersTable::new(43)),
})
.await
.unwrap_err();
assert!(
err.to_string()
.contains("Table `greptime.public.test_table` already exists"),
"Actual error message: {err}",
);
}
#[test]
fn test_concurrent_register() {
common_telemetry::init_default_ut_logging();
let rt = Arc::new(tokio::runtime::Builder::new_multi_thread().build().unwrap());
let (_dir, catalog_manager) =
rt.block_on(async { create_local_catalog_manager().await.unwrap() });
let catalog_manager = Arc::new(catalog_manager);
let succeed: Arc<Mutex<Option<TableRef>>> = Arc::new(Mutex::new(None));
let mut handles = Vec::with_capacity(8);
for i in 0..8 {
let catalog = catalog_manager.clone();
let succeed = succeed.clone();
let handle = rt.spawn(async move {
let table_id = 42 + i;
let table = Arc::new(NumbersTable::new(table_id));
let req = RegisterTableRequest {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table_name: "test_table".to_string(),
table_id,
table: table.clone(),
};
match catalog.register_table(req).await {
Ok(res) => {
if res {
let mut succeed = succeed.lock().await;
info!("Successfully registered table: {}", table_id);
*succeed = Some(table);
}
}
Err(_) => {
error!("Failed to register table {}", table_id);
}
}
});
handles.push(handle);
}
rt.block_on(async move {
for handle in handles {
handle.await.unwrap();
}
let guard = succeed.lock().await;
let table = guard.as_ref().unwrap();
let table_registered = catalog_manager
.table(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, "test_table")
.await
.unwrap()
.unwrap();
assert_eq!(
table_registered.table_info().ident.table_id,
table.table_info().ident.table_id
);
});
}
}

View File

@@ -1,490 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#![feature(assert_matches)]
#[cfg(test)]
mod tests {
use std::assert_matches::assert_matches;
use std::collections::HashSet;
use std::sync::Arc;
use std::time::Duration;
use catalog::error::Error;
use catalog::remote::mock::MockTableEngine;
use catalog::remote::region_alive_keeper::RegionAliveKeepers;
use catalog::remote::{CachedMetaKvBackend, RemoteCatalogManager};
use catalog::{CatalogManager, RegisterSchemaRequest, RegisterTableRequest};
use common_catalog::consts::{
DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, INFORMATION_SCHEMA_NAME, MITO_ENGINE,
};
use common_meta::helper::{CatalogKey, CatalogValue, SchemaKey, SchemaValue};
use common_meta::ident::TableIdent;
use common_meta::key::TableMetadataManager;
use common_meta::kv_backend::memory::MemoryKvBackend;
use common_meta::kv_backend::KvBackend;
use common_meta::rpc::store::{CompareAndPutRequest, PutRequest, RangeRequest};
use datatypes::schema::RawSchema;
use table::engine::manager::{MemoryTableEngineManager, TableEngineManagerRef};
use table::engine::{EngineContext, TableEngineRef};
use table::requests::CreateTableRequest;
use table::test_util::EmptyTable;
use tokio::time::Instant;
struct TestingComponents {
catalog_manager: Arc<RemoteCatalogManager>,
table_engine_manager: TableEngineManagerRef,
region_alive_keepers: Arc<RegionAliveKeepers>,
}
impl TestingComponents {
fn table_engine(&self) -> TableEngineRef {
self.table_engine_manager.engine(MITO_ENGINE).unwrap()
}
}
#[tokio::test]
async fn test_backend() {
let backend = MemoryKvBackend::<Error>::default();
let default_catalog_key = CatalogKey {
catalog_name: DEFAULT_CATALOG_NAME.to_string(),
}
.to_string();
let req = PutRequest::new()
.with_key(default_catalog_key.as_bytes())
.with_value(CatalogValue.as_bytes().unwrap());
backend.put(req).await.unwrap();
let schema_key = SchemaKey {
catalog_name: DEFAULT_CATALOG_NAME.to_string(),
schema_name: DEFAULT_SCHEMA_NAME.to_string(),
}
.to_string();
let req = PutRequest::new()
.with_key(schema_key.as_bytes())
.with_value(SchemaValue.as_bytes().unwrap());
backend.put(req).await.unwrap();
let req = RangeRequest::new().with_prefix(b"__c-".to_vec());
let res = backend
.range(req)
.await
.unwrap()
.kvs
.into_iter()
.map(|kv| String::from_utf8_lossy(kv.key()).to_string());
assert_eq!(
vec!["__c-greptime".to_string()],
res.into_iter().collect::<Vec<_>>()
);
}
#[tokio::test]
async fn test_cached_backend() {
let backend = CachedMetaKvBackend::wrap(Arc::new(MemoryKvBackend::default()));
let default_catalog_key = CatalogKey {
catalog_name: DEFAULT_CATALOG_NAME.to_string(),
}
.to_string();
let req = PutRequest::new()
.with_key(default_catalog_key.as_bytes())
.with_value(CatalogValue.as_bytes().unwrap());
backend.put(req).await.unwrap();
let ret = backend.get(b"__c-greptime").await.unwrap();
let _ = ret.unwrap();
let req = CompareAndPutRequest::new()
.with_key(b"__c-greptime".to_vec())
.with_expect(CatalogValue.as_bytes().unwrap())
.with_value(b"123".to_vec());
let _ = backend.compare_and_put(req).await.unwrap();
let ret = backend.get(b"__c-greptime").await.unwrap();
assert_eq!(b"123", ret.as_ref().unwrap().value.as_slice());
let req = PutRequest::new()
.with_key(b"__c-greptime".to_vec())
.with_value(b"1234".to_vec());
let _ = backend.put(req).await;
let ret = backend.get(b"__c-greptime").await.unwrap();
assert_eq!(b"1234", ret.unwrap().value.as_slice());
backend.delete(b"__c-greptime", false).await.unwrap();
let ret = backend.get(b"__c-greptime").await.unwrap();
assert!(ret.is_none());
}
async fn prepare_components(node_id: u64) -> TestingComponents {
let backend = Arc::new(MemoryKvBackend::default());
let req = PutRequest::new()
.with_key(b"__c-greptime".to_vec())
.with_value(b"".to_vec());
backend.put(req).await.unwrap();
let req = PutRequest::new()
.with_key(b"__s-greptime-public".to_vec())
.with_value(b"".to_vec());
backend.put(req).await.unwrap();
let cached_backend = Arc::new(CachedMetaKvBackend::wrap(backend));
let table_engine = Arc::new(MockTableEngine::default());
let engine_manager = Arc::new(MemoryTableEngineManager::alias(
MITO_ENGINE.to_string(),
table_engine,
));
let region_alive_keepers = Arc::new(RegionAliveKeepers::new(engine_manager.clone(), 5000));
let catalog_manager = RemoteCatalogManager::new(
engine_manager.clone(),
node_id,
cached_backend.clone(),
region_alive_keepers.clone(),
Arc::new(TableMetadataManager::new(cached_backend)),
);
catalog_manager.start().await.unwrap();
TestingComponents {
catalog_manager: Arc::new(catalog_manager),
table_engine_manager: engine_manager,
region_alive_keepers,
}
}
#[tokio::test]
async fn test_remote_catalog_default() {
common_telemetry::init_default_ut_logging();
let node_id = 42;
let TestingComponents {
catalog_manager, ..
} = prepare_components(node_id).await;
assert_eq!(
vec![DEFAULT_CATALOG_NAME.to_string()],
catalog_manager.catalog_names().await.unwrap()
);
let mut schema_names = catalog_manager
.schema_names(DEFAULT_CATALOG_NAME)
.await
.unwrap();
schema_names.sort_unstable();
assert_eq!(
vec![
INFORMATION_SCHEMA_NAME.to_string(),
DEFAULT_SCHEMA_NAME.to_string()
],
schema_names
);
}
#[tokio::test]
async fn test_remote_catalog_register_nonexistent() {
common_telemetry::init_default_ut_logging();
let node_id = 42;
let components = prepare_components(node_id).await;
// register a new table with an nonexistent catalog
let catalog_name = "nonexistent_catalog".to_string();
let schema_name = "nonexistent_schema".to_string();
let table_name = "fail_table".to_string();
// this schema has no effect
let table_schema = RawSchema::new(vec![]);
let table = components
.table_engine()
.create_table(
&EngineContext {},
CreateTableRequest {
id: 1,
catalog_name: catalog_name.clone(),
schema_name: schema_name.clone(),
table_name: table_name.clone(),
desc: None,
schema: table_schema,
region_numbers: vec![0],
primary_key_indices: vec![],
create_if_not_exists: false,
table_options: Default::default(),
engine: MITO_ENGINE.to_string(),
},
)
.await
.unwrap();
let reg_req = RegisterTableRequest {
catalog: catalog_name,
schema: schema_name,
table_name,
table_id: 1,
table,
};
let res = components.catalog_manager.register_table(reg_req).await;
// because nonexistent_catalog does not exist yet.
assert_matches!(
res.err().unwrap(),
catalog::error::Error::CatalogNotFound { .. }
);
}
#[tokio::test]
async fn test_register_table() {
let node_id = 42;
let components = prepare_components(node_id).await;
let mut schema_names = components
.catalog_manager
.schema_names(DEFAULT_CATALOG_NAME)
.await
.unwrap();
schema_names.sort_unstable();
assert_eq!(
vec![
INFORMATION_SCHEMA_NAME.to_string(),
DEFAULT_SCHEMA_NAME.to_string(),
],
schema_names
);
// register a new table with an nonexistent catalog
let catalog_name = DEFAULT_CATALOG_NAME.to_string();
let schema_name = DEFAULT_SCHEMA_NAME.to_string();
let table_name = "test_table".to_string();
let table_id = 1;
// this schema has no effect
let table_schema = RawSchema::new(vec![]);
let table = components
.table_engine()
.create_table(
&EngineContext {},
CreateTableRequest {
id: table_id,
catalog_name: catalog_name.clone(),
schema_name: schema_name.clone(),
table_name: table_name.clone(),
desc: None,
schema: table_schema,
region_numbers: vec![0],
primary_key_indices: vec![],
create_if_not_exists: false,
table_options: Default::default(),
engine: MITO_ENGINE.to_string(),
},
)
.await
.unwrap();
let reg_req = RegisterTableRequest {
catalog: catalog_name,
schema: schema_name,
table_name: table_name.clone(),
table_id,
table,
};
assert!(components
.catalog_manager
.register_table(reg_req)
.await
.unwrap());
assert_eq!(
vec![table_name],
components
.catalog_manager
.table_names(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME)
.await
.unwrap()
);
}
#[tokio::test]
async fn test_register_catalog_schema_table() {
let node_id = 42;
let components = prepare_components(node_id).await;
let catalog_name = "test_catalog".to_string();
let schema_name = "nonexistent_schema".to_string();
// register catalog to catalog manager
assert!(components
.catalog_manager
.clone()
.register_catalog(catalog_name.clone())
.await
.is_ok());
assert_eq!(
HashSet::<String>::from_iter(vec![
DEFAULT_CATALOG_NAME.to_string(),
catalog_name.clone()
]),
HashSet::from_iter(components.catalog_manager.catalog_names().await.unwrap())
);
let table_to_register = components
.table_engine()
.create_table(
&EngineContext {},
CreateTableRequest {
id: 2,
catalog_name: catalog_name.clone(),
schema_name: schema_name.clone(),
table_name: "".to_string(),
desc: None,
schema: RawSchema::new(vec![]),
region_numbers: vec![0],
primary_key_indices: vec![],
create_if_not_exists: false,
table_options: Default::default(),
engine: MITO_ENGINE.to_string(),
},
)
.await
.unwrap();
let reg_req = RegisterTableRequest {
catalog: catalog_name.clone(),
schema: schema_name.clone(),
table_name: " fail_table".to_string(),
table_id: 2,
table: table_to_register,
};
// this register will fail since schema does not exist yet
assert_matches!(
components
.catalog_manager
.register_table(reg_req.clone())
.await
.unwrap_err(),
catalog::error::Error::SchemaNotFound { .. }
);
let register_schema_request = RegisterSchemaRequest {
catalog: catalog_name.to_string(),
schema: schema_name.to_string(),
};
assert!(components
.catalog_manager
.register_schema(register_schema_request)
.await
.expect("Register schema should not fail"));
assert!(components
.catalog_manager
.register_table(reg_req)
.await
.unwrap());
assert_eq!(
HashSet::from([schema_name.clone(), INFORMATION_SCHEMA_NAME.to_string()]),
components
.catalog_manager
.schema_names(&catalog_name)
.await
.unwrap()
.into_iter()
.collect()
)
}
#[tokio::test]
async fn test_register_table_before_and_after_region_alive_keeper_started() {
let components = prepare_components(42).await;
let catalog_manager = &components.catalog_manager;
let region_alive_keepers = &components.region_alive_keepers;
let table_before = TableIdent {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table: "table_before".to_string(),
table_id: 1,
engine: MITO_ENGINE.to_string(),
};
let request = RegisterTableRequest {
catalog: table_before.catalog.clone(),
schema: table_before.schema.clone(),
table_name: table_before.table.clone(),
table_id: table_before.table_id,
table: Arc::new(EmptyTable::new(CreateTableRequest {
id: table_before.table_id,
catalog_name: table_before.catalog.clone(),
schema_name: table_before.schema.clone(),
table_name: table_before.table.clone(),
desc: None,
schema: RawSchema::new(vec![]),
region_numbers: vec![0],
primary_key_indices: vec![],
create_if_not_exists: false,
table_options: Default::default(),
engine: MITO_ENGINE.to_string(),
})),
};
assert!(catalog_manager.register_table(request).await.unwrap());
let keeper = region_alive_keepers
.find_keeper(&table_before)
.await
.unwrap();
let deadline = keeper.deadline(0).await.unwrap();
let far_future = Instant::now() + Duration::from_secs(86400 * 365 * 29);
// assert region alive countdown is not started
assert!(deadline > far_future);
region_alive_keepers.start().await;
let table_after = TableIdent {
catalog: DEFAULT_CATALOG_NAME.to_string(),
schema: DEFAULT_SCHEMA_NAME.to_string(),
table: "table_after".to_string(),
table_id: 2,
engine: MITO_ENGINE.to_string(),
};
let request = RegisterTableRequest {
catalog: table_after.catalog.clone(),
schema: table_after.schema.clone(),
table_name: table_after.table.clone(),
table_id: table_after.table_id,
table: Arc::new(EmptyTable::new(CreateTableRequest {
id: table_after.table_id,
catalog_name: table_after.catalog.clone(),
schema_name: table_after.schema.clone(),
table_name: table_after.table.clone(),
desc: None,
schema: RawSchema::new(vec![]),
region_numbers: vec![0],
primary_key_indices: vec![],
create_if_not_exists: false,
table_options: Default::default(),
engine: MITO_ENGINE.to_string(),
})),
};
assert!(catalog_manager.register_table(request).await.unwrap());
let keeper = region_alive_keepers
.find_keeper(&table_after)
.await
.unwrap();
let deadline = keeper.deadline(0).await.unwrap();
// assert countdown is started for the table registered after [RegionAliveKeepers] started
assert!(deadline <= Instant::now() + Duration::from_secs(20));
let keeper = region_alive_keepers
.find_keeper(&table_before)
.await
.unwrap();
let deadline = keeper.deadline(0).await.unwrap();
// assert countdown is started for the table registered before [RegionAliveKeepers] started, too
assert!(deadline <= Instant::now() + Duration::from_secs(20));
}
}

View File

@@ -11,10 +11,12 @@ testing = []
api = { workspace = true }
arrow-flight.workspace = true
async-stream.workspace = true
async-trait.workspace = true
common-base = { workspace = true }
common-catalog = { workspace = true }
common-error = { workspace = true }
common-grpc = { workspace = true }
common-macro = { workspace = true }
common-meta = { workspace = true }
common-query = { workspace = true }
common-recordbatch = { workspace = true }
@@ -22,12 +24,14 @@ common-telemetry = { workspace = true }
common-time = { workspace = true }
datafusion.workspace = true
datatypes = { workspace = true }
derive_builder.workspace = true
enum_dispatch = "0.3"
futures-util.workspace = true
moka = { version = "0.9", features = ["future"] }
moka = { workspace = true, features = ["future"] }
parking_lot = "0.12"
prost.workspace = true
rand.workspace = true
session = { workspace = true }
snafu.workspace = true
tokio-stream = { version = "0.1", features = ["net"] }
tokio.workspace = true

View File

@@ -12,7 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
use api::v1::{ColumnDataType, ColumnDef, CreateTableExpr, TableId};
use api::v1::{ColumnDataType, ColumnDef, CreateTableExpr, SemanticType, TableId};
use client::{Client, Database};
use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, MITO_ENGINE};
use prost::Message;
@@ -41,21 +41,27 @@ async fn run() {
column_defs: vec![
ColumnDef {
name: "timestamp".to_string(),
datatype: ColumnDataType::TimestampMillisecond as i32,
data_type: ColumnDataType::TimestampMillisecond as i32,
is_nullable: false,
default_constraint: vec![],
semantic_type: SemanticType::Timestamp as i32,
comment: String::new(),
},
ColumnDef {
name: "key".to_string(),
datatype: ColumnDataType::Uint64 as i32,
data_type: ColumnDataType::Uint64 as i32,
is_nullable: false,
default_constraint: vec![],
semantic_type: SemanticType::Tag as i32,
comment: String::new(),
},
ColumnDef {
name: "value".to_string(),
datatype: ColumnDataType::Uint64 as i32,
data_type: ColumnDataType::Uint64 as i32,
is_nullable: false,
default_constraint: vec![],
semantic_type: SemanticType::Field as i32,
comment: String::new(),
},
],
time_index: "timestamp".to_string(),
@@ -63,7 +69,6 @@ async fn run() {
create_if_not_exists: false,
table_options: Default::default(),
table_id: Some(TableId { id: 1024 }),
region_numbers: vec![0],
engine: MITO_ENGINE.to_string(),
};
@@ -73,7 +78,7 @@ async fn run() {
let logical = mock_logical_plan();
event!(Level::INFO, "plan size: {:#?}", logical.len());
let result = db.logical_plan(logical, None).await.unwrap();
let result = db.logical_plan(logical, 0).await.unwrap();
event!(Level::INFO, "result: {:#?}", result);
}

View File

@@ -42,14 +42,14 @@ async fn run() {
.insert(vec![to_insert_request(weather_records_1())])
.await
{
error!("Error: {e}");
error!("Error: {e:?}");
}
if let Err(e) = stream_inserter
.insert(vec![to_insert_request(weather_records_2())])
.await
{
error!("Error: {e}");
error!("Error: {e:?}");
}
let result = stream_inserter.finish().await;
@@ -59,7 +59,7 @@ async fn run() {
info!("Rows written: {rows}");
}
Err(e) => {
error!("Error: {e}");
error!("Error: {e:?}");
}
};
}
@@ -131,7 +131,7 @@ fn to_insert_request(records: Vec<WeatherRecord>) -> InsertRequest {
Column {
column_name: "ts".to_owned(),
values: Some(column::Values {
ts_millisecond_values: timestamp_millis,
timestamp_millisecond_values: timestamp_millis,
..Default::default()
}),
semantic_type: SemanticType::Timestamp as i32,
@@ -177,6 +177,5 @@ fn to_insert_request(records: Vec<WeatherRecord>) -> InsertRequest {
table_name: "weather_demo".to_owned(),
columns,
row_count: rows as u32,
..Default::default()
}
}

View File

@@ -17,6 +17,7 @@ use std::sync::Arc;
use api::v1::greptime_database_client::GreptimeDatabaseClient;
use api::v1::health_check_client::HealthCheckClient;
use api::v1::prometheus_gateway_client::PrometheusGatewayClient;
use api::v1::region::region_client::RegionClient as PbRegionClient;
use api::v1::HealthCheckRequest;
use arrow_flight::flight_service_client::FlightServiceClient;
use common_grpc::channel_manager::ChannelManager;
@@ -82,11 +83,6 @@ impl Client {
Default::default()
}
pub fn with_manager(channel_manager: ChannelManager) -> Self {
let inner = Arc::new(Inner::with_manager(channel_manager));
Self { inner }
}
pub fn with_urls<U, A>(urls: A) -> Self
where
U: AsRef<str>,
@@ -142,21 +138,48 @@ impl Client {
Ok((addr, channel))
}
fn max_grpc_recv_message_size(&self) -> usize {
self.inner
.channel_manager
.config()
.max_recv_message_size
.as_bytes() as usize
}
fn max_grpc_send_message_size(&self) -> usize {
self.inner
.channel_manager
.config()
.max_send_message_size
.as_bytes() as usize
}
pub(crate) fn make_flight_client(&self) -> Result<FlightClient> {
let (addr, channel) = self.find_channel()?;
Ok(FlightClient {
addr,
client: FlightServiceClient::new(channel),
client: FlightServiceClient::new(channel)
.max_decoding_message_size(self.max_grpc_recv_message_size())
.max_encoding_message_size(self.max_grpc_send_message_size()),
})
}
pub(crate) fn make_database_client(&self) -> Result<DatabaseClient> {
let (_, channel) = self.find_channel()?;
Ok(DatabaseClient {
inner: GreptimeDatabaseClient::new(channel),
inner: GreptimeDatabaseClient::new(channel)
.max_decoding_message_size(self.max_grpc_recv_message_size())
.max_encoding_message_size(self.max_grpc_send_message_size()),
})
}
pub(crate) fn raw_region_client(&self) -> Result<PbRegionClient<Channel>> {
let (_, channel) = self.find_channel()?;
Ok(PbRegionClient::new(channel)
.max_decoding_message_size(self.max_grpc_recv_message_size())
.max_encoding_message_size(self.max_grpc_send_message_size()))
}
pub fn make_prometheus_gateway_client(&self) -> Result<PrometheusGatewayClient<Channel>> {
let (_, channel) = self.find_channel()?;
Ok(PrometheusGatewayClient::new(channel))

View File

@@ -13,12 +13,15 @@
// limitations under the License.
use std::fmt::{Debug, Formatter};
use std::sync::Arc;
use std::time::Duration;
use common_grpc::channel_manager::{ChannelConfig, ChannelManager};
use common_meta::datanode_manager::{Datanode, DatanodeManager};
use common_meta::peer::Peer;
use moka::future::{Cache, CacheBuilder};
use crate::region::RegionRequester;
use crate::Client;
pub struct DatanodeClients {
@@ -40,6 +43,15 @@ impl Debug for DatanodeClients {
}
}
#[async_trait::async_trait]
impl DatanodeManager for DatanodeClients {
async fn datanode(&self, datanode: &Peer) -> Arc<dyn Datanode> {
let client = self.get_client(datanode).await;
Arc::new(RegionRequester::new(client))
}
}
impl DatanodeClients {
pub fn new(config: ChannelConfig) -> Self {
Self {

View File

@@ -17,20 +17,23 @@ use api::v1::ddl_request::Expr as DdlExpr;
use api::v1::greptime_request::Request;
use api::v1::query_request::Query;
use api::v1::{
AlterExpr, AuthHeader, CompactTableExpr, CreateTableExpr, DdlRequest, DeleteRequest,
DropTableExpr, FlushTableExpr, GreptimeRequest, InsertRequests, PromRangeQuery, QueryRequest,
RequestHeader, TruncateTableExpr,
AlterExpr, AuthHeader, CreateTableExpr, DdlRequest, DeleteRequests, DropTableExpr,
GreptimeRequest, InsertRequests, PromRangeQuery, QueryRequest, RequestHeader,
RowInsertRequests, TruncateTableExpr,
};
use arrow_flight::{FlightData, Ticket};
use arrow_flight::Ticket;
use async_stream::stream;
use common_error::ext::{BoxedError, ErrorExt};
use common_grpc::flight::{flight_messages_to_recordbatches, FlightDecoder, FlightMessage};
use common_grpc::flight::{FlightDecoder, FlightMessage};
use common_query::Output;
use common_recordbatch::error::ExternalSnafu;
use common_recordbatch::RecordBatchStreamAdaptor;
use common_telemetry::{logging, timer};
use futures_util::{TryFutureExt, TryStreamExt};
use futures_util::StreamExt;
use prost::Message;
use snafu::{ensure, ResultExt};
use crate::error::{ConvertFlightDataSnafu, IllegalFlightMessagesSnafu, ServerSnafu};
use crate::error::{ConvertFlightDataSnafu, Error, IllegalFlightMessagesSnafu, ServerSnafu};
use crate::{error, from_grpc_response, metrics, Client, Result, StreamInserter};
#[derive(Clone, Debug, Default)]
@@ -112,6 +115,11 @@ impl Database {
self.handle(Request::Inserts(requests)).await
}
pub async fn row_insert(&self, requests: RowInsertRequests) -> Result<u32> {
let _timer = timer!(metrics::METRIC_GRPC_INSERT);
self.handle(Request::RowInserts(requests)).await
}
pub fn streaming_inserter(&self) -> Result<StreamInserter> {
self.streaming_inserter_with_channel_size(65536)
}
@@ -132,20 +140,20 @@ impl Database {
Ok(stream_inserter)
}
pub async fn delete(&self, request: DeleteRequest) -> Result<u32> {
pub async fn delete(&self, request: DeleteRequests) -> Result<u32> {
let _timer = timer!(metrics::METRIC_GRPC_DELETE);
self.handle(Request::Delete(request)).await
self.handle(Request::Deletes(request)).await
}
async fn handle(&self, request: Request) -> Result<u32> {
let mut client = self.client.make_database_client()?.inner;
let request = self.to_rpc_request(request, None);
let request = self.to_rpc_request(request, 0);
let response = client.handle(request).await?.into_inner();
from_grpc_response(response)
}
#[inline]
fn to_rpc_request(&self, request: Request, trace_id: Option<u64>) -> GreptimeRequest {
fn to_rpc_request(&self, request: Request, trace_id: u64) -> GreptimeRequest {
GreptimeRequest {
header: Some(RequestHeader {
catalog: self.catalog.clone(),
@@ -153,7 +161,7 @@ impl Database {
authorization: self.ctx.auth_header.clone(),
dbname: self.dbname.clone(),
trace_id,
span_id: None,
span_id: 0,
}),
request: Some(request),
}
@@ -165,16 +173,12 @@ impl Database {
Request::Query(QueryRequest {
query: Some(Query::Sql(sql.to_string())),
}),
None,
0,
)
.await
}
pub async fn logical_plan(
&self,
logical_plan: Vec<u8>,
trace_id: Option<u64>,
) -> Result<Output> {
pub async fn logical_plan(&self, logical_plan: Vec<u8>, trace_id: u64) -> Result<Output> {
let _timer = timer!(metrics::METRIC_GRPC_LOGICAL_PLAN);
self.do_get(
Request::Query(QueryRequest {
@@ -202,7 +206,7 @@ impl Database {
step: step.to_string(),
})),
}),
None,
0,
)
.await
}
@@ -213,7 +217,7 @@ impl Database {
Request::Ddl(DdlRequest {
expr: Some(DdlExpr::CreateTable(expr)),
}),
None,
0,
)
.await
}
@@ -224,7 +228,7 @@ impl Database {
Request::Ddl(DdlRequest {
expr: Some(DdlExpr::Alter(expr)),
}),
None,
0,
)
.await
}
@@ -235,29 +239,7 @@ impl Database {
Request::Ddl(DdlRequest {
expr: Some(DdlExpr::DropTable(expr)),
}),
None,
)
.await
}
pub async fn flush_table(&self, expr: FlushTableExpr) -> Result<Output> {
let _timer = timer!(metrics::METRIC_GRPC_FLUSH_TABLE);
self.do_get(
Request::Ddl(DdlRequest {
expr: Some(DdlExpr::FlushTable(expr)),
}),
None,
)
.await
}
pub async fn compact_table(&self, expr: CompactTableExpr) -> Result<Output> {
let _timer = timer!(metrics::METRIC_GRPC_COMPACT_TABLE);
self.do_get(
Request::Ddl(DdlRequest {
expr: Some(DdlExpr::CompactTable(expr)),
}),
None,
0,
)
.await
}
@@ -268,12 +250,12 @@ impl Database {
Request::Ddl(DdlRequest {
expr: Some(DdlExpr::TruncateTable(expr)),
}),
None,
0,
)
.await
}
async fn do_get(&self, request: Request, trace_id: Option<u64>) -> Result<Output> {
async fn do_get(&self, request: Request, trace_id: u64) -> Result<Output> {
// FIXME(paomian): should be added some labels for metrics
let _timer = timer!(metrics::METRIC_GRPC_DO_GET);
let request = self.to_rpc_request(request, trace_id);
@@ -283,55 +265,81 @@ impl Database {
let mut client = self.client.make_flight_client()?;
let flight_data: Vec<FlightData> = client
.mut_inner()
.do_get(request)
.and_then(|response| response.into_inner().try_collect())
.await
.map_err(|e| {
let tonic_code = e.code();
let e: error::Error = e.into();
let code = e.status_code();
let msg = e.to_string();
ServerSnafu { code, msg }
.fail::<()>()
.map_err(BoxedError::new)
.context(error::FlightGetSnafu {
tonic_code,
addr: client.addr(),
})
.map_err(|error| {
logging::error!(
"Failed to do Flight get, addr: {}, code: {}, source: {}",
client.addr(),
tonic_code,
error
);
error
})
.unwrap_err()
})?;
let decoder = &mut FlightDecoder::default();
let flight_messages = flight_data
.into_iter()
.map(|x| decoder.try_decode(x).context(ConvertFlightDataSnafu))
.collect::<Result<Vec<_>>>()?;
let output = if let Some(FlightMessage::AffectedRows(rows)) = flight_messages.get(0) {
ensure!(
flight_messages.len() == 1,
IllegalFlightMessagesSnafu {
reason: "Expect 'AffectedRows' Flight messages to be one and only!"
}
let response = client.mut_inner().do_get(request).await.map_err(|e| {
let tonic_code = e.code();
let e: error::Error = e.into();
let code = e.status_code();
let msg = e.to_string();
let error = Error::FlightGet {
tonic_code,
addr: client.addr().to_string(),
source: BoxedError::new(ServerSnafu { code, msg }.build()),
};
logging::error!(
"Failed to do Flight get, addr: {}, code: {}, source: {:?}",
client.addr(),
tonic_code,
error
);
Output::AffectedRows(*rows)
} else {
let recordbatches = flight_messages_to_recordbatches(flight_messages)
.context(ConvertFlightDataSnafu)?;
Output::RecordBatches(recordbatches)
error
})?;
let flight_data_stream = response.into_inner();
let mut decoder = FlightDecoder::default();
let mut flight_message_stream = flight_data_stream.map(move |flight_data| {
flight_data
.map_err(Error::from)
.and_then(|data| decoder.try_decode(data).context(ConvertFlightDataSnafu))
});
let Some(first_flight_message) = flight_message_stream.next().await else {
return IllegalFlightMessagesSnafu {
reason: "Expect the response not to be empty",
}
.fail();
};
Ok(output)
let first_flight_message = first_flight_message?;
match first_flight_message {
FlightMessage::AffectedRows(rows) => {
ensure!(
flight_message_stream.next().await.is_none(),
IllegalFlightMessagesSnafu {
reason: "Expect 'AffectedRows' Flight messages to be the one and the only!"
}
);
Ok(Output::AffectedRows(rows))
}
FlightMessage::Recordbatch(_) => IllegalFlightMessagesSnafu {
reason: "The first flight message cannot be a RecordBatch message",
}
.fail(),
FlightMessage::Schema(schema) => {
let stream = Box::pin(stream!({
while let Some(flight_message) = flight_message_stream.next().await {
let flight_message = flight_message
.map_err(BoxedError::new)
.context(ExternalSnafu)?;
let FlightMessage::Recordbatch(record_batch) = flight_message else {
yield IllegalFlightMessagesSnafu {reason: "A Schema message must be succeeded exclusively by a set of RecordBatch messages"}
.fail()
.map_err(BoxedError::new)
.context(ExternalSnafu);
break;
};
yield Ok(record_batch);
}
}));
let record_batch_stream = RecordBatchStreamAdaptor {
schema,
stream,
output_ordering: None,
};
Ok(Output::Stream(Box::pin(record_batch_stream)))
}
}
}
}
@@ -342,106 +350,11 @@ pub struct FlightContext {
#[cfg(test)]
mod tests {
use std::sync::Arc;
use api::helper::ColumnDataTypeWrapper;
use api::v1::auth_header::AuthScheme;
use api::v1::{AuthHeader, Basic, Column};
use common_grpc::select::{null_mask, values};
use common_grpc_expr::column_to_vector;
use datatypes::prelude::{Vector, VectorRef};
use datatypes::vectors::{
BinaryVector, BooleanVector, DateTimeVector, DateVector, Float32Vector, Float64Vector,
Int16Vector, Int32Vector, Int64Vector, Int8Vector, StringVector, UInt16Vector,
UInt32Vector, UInt64Vector, UInt8Vector,
};
use api::v1::{AuthHeader, Basic};
use crate::database::FlightContext;
#[test]
fn test_column_to_vector() {
let mut column = create_test_column(Arc::new(BooleanVector::from(vec![true])));
column.datatype = -100;
let result = column_to_vector(&column, 1);
assert!(result.is_err());
assert_eq!(
result.unwrap_err().to_string(),
"Column datatype error, source: Unknown proto column datatype: -100"
);
macro_rules! test_with_vector {
($vector: expr) => {
let vector = Arc::new($vector);
let column = create_test_column(vector.clone());
let result = column_to_vector(&column, vector.len() as u32).unwrap();
assert_eq!(result, vector as VectorRef);
};
}
test_with_vector!(BooleanVector::from(vec![Some(true), None, Some(false)]));
test_with_vector!(Int8Vector::from(vec![Some(i8::MIN), None, Some(i8::MAX)]));
test_with_vector!(Int16Vector::from(vec![
Some(i16::MIN),
None,
Some(i16::MAX)
]));
test_with_vector!(Int32Vector::from(vec![
Some(i32::MIN),
None,
Some(i32::MAX)
]));
test_with_vector!(Int64Vector::from(vec![
Some(i64::MIN),
None,
Some(i64::MAX)
]));
test_with_vector!(UInt8Vector::from(vec![Some(u8::MIN), None, Some(u8::MAX)]));
test_with_vector!(UInt16Vector::from(vec![
Some(u16::MIN),
None,
Some(u16::MAX)
]));
test_with_vector!(UInt32Vector::from(vec![
Some(u32::MIN),
None,
Some(u32::MAX)
]));
test_with_vector!(UInt64Vector::from(vec![
Some(u64::MIN),
None,
Some(u64::MAX)
]));
test_with_vector!(Float32Vector::from(vec![
Some(f32::MIN),
None,
Some(f32::MAX)
]));
test_with_vector!(Float64Vector::from(vec![
Some(f64::MIN),
None,
Some(f64::MAX)
]));
test_with_vector!(BinaryVector::from(vec![
Some(b"".to_vec()),
None,
Some(b"hello".to_vec())
]));
test_with_vector!(StringVector::from(vec![Some(""), None, Some("foo"),]));
test_with_vector!(DateVector::from(vec![Some(1), None, Some(3)]));
test_with_vector!(DateTimeVector::from(vec![Some(4), None, Some(6)]));
}
fn create_test_column(vector: VectorRef) -> Column {
let wrapper: ColumnDataTypeWrapper = vector.data_type().try_into().unwrap();
Column {
column_name: "test".to_string(),
semantic_type: 1,
values: Some(values(&[vector.clone()]).unwrap()),
null_mask: null_mask(&[vector.clone()], vector.len()),
datatype: wrapper.datatype() as i32,
}
}
#[test]
fn test_flight_ctx() {
let mut ctx = FlightContext::default();

View File

@@ -13,34 +13,41 @@
// limitations under the License.
use std::any::Any;
use std::str::FromStr;
use common_error::ext::{BoxedError, ErrorExt};
use common_error::status_code::StatusCode;
use common_error::{INNER_ERROR_CODE, INNER_ERROR_MSG};
use common_error::{GREPTIME_ERROR_CODE, GREPTIME_ERROR_MSG};
use common_macro::stack_trace_debug;
use snafu::{Location, Snafu};
use tonic::{Code, Status};
#[derive(Debug, Snafu)]
#[derive(Snafu)]
#[snafu(visibility(pub))]
#[stack_trace_debug]
pub enum Error {
#[snafu(display("Illegal Flight messages, reason: {}", reason))]
IllegalFlightMessages { reason: String, location: Location },
#[snafu(display("Failed to do Flight get, code: {}, source: {}", tonic_code, source))]
#[snafu(display("Failed to do Flight get, code: {}", tonic_code))]
FlightGet {
addr: String,
tonic_code: Code,
source: BoxedError,
},
#[snafu(display("Failed to convert FlightData, source: {}", source))]
#[snafu(display("Failure occurs during handling request"))]
HandleRequest {
location: Location,
source: BoxedError,
},
#[snafu(display("Failed to convert FlightData"))]
ConvertFlightData {
location: Location,
source: common_grpc::Error,
},
#[snafu(display("Column datatype error, source: {}", source))]
#[snafu(display("Column datatype error"))]
ColumnDataType {
location: Location,
source: api::error::Error,
@@ -52,17 +59,16 @@ pub enum Error {
#[snafu(display("Missing required field in protobuf, field: {}", field))]
MissingField { field: String, location: Location },
#[snafu(display(
"Failed to create gRPC channel, peer address: {}, source: {}",
addr,
source
))]
#[snafu(display("Failed to create gRPC channel, peer address: {}", addr))]
CreateChannel {
addr: String,
location: Location,
source: common_grpc::error::Error,
},
#[snafu(display("Failed to request RegionServer, code: {}", code))]
RegionServer { code: Code, source: BoxedError },
// Server error carried in Tonic Status's metadata.
#[snafu(display("{}", msg))]
Server { code: StatusCode, msg: String },
@@ -86,7 +92,9 @@ impl ErrorExt for Error {
| Error::ClientStreaming { .. } => StatusCode::Internal,
Error::Server { code, .. } => *code,
Error::FlightGet { source, .. } => source.status_code(),
Error::FlightGet { source, .. }
| Error::HandleRequest { source, .. }
| Error::RegionServer { source, .. } => source.status_code(),
Error::CreateChannel { source, .. } | Error::ConvertFlightData { source, .. } => {
source.status_code()
}
@@ -107,11 +115,18 @@ impl From<Status> for Error {
.and_then(|v| String::from_utf8(v.as_bytes().to_vec()).ok())
}
let code = get_metadata_value(&e, INNER_ERROR_CODE)
.and_then(|s| StatusCode::from_str(&s).ok())
let code = get_metadata_value(&e, GREPTIME_ERROR_CODE)
.and_then(|s| {
if let Ok(code) = s.parse::<u32>() {
StatusCode::from_u32(code)
} else {
None
}
})
.unwrap_or(StatusCode::Unknown);
let msg = get_metadata_value(&e, INNER_ERROR_MSG).unwrap_or(e.to_string());
let msg =
get_metadata_value(&e, GREPTIME_ERROR_MSG).unwrap_or_else(|| e.message().to_string());
Self::Server { code, msg }
}

View File

@@ -18,6 +18,7 @@ mod database;
pub mod error;
pub mod load_balance;
mod metrics;
pub mod region;
mod stream_insert;
pub use api;
@@ -25,6 +26,8 @@ use api::v1::greptime_response::Response;
use api::v1::{AffectedRows, GreptimeResponse};
pub use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
use common_error::status_code::StatusCode;
pub use common_query::Output;
pub use common_recordbatch::{RecordBatches, SendableRecordBatchStream};
use snafu::OptionExt;
pub use self::client::Client;

View File

@@ -21,7 +21,6 @@ pub const METRIC_GRPC_SQL: &str = "grpc.sql";
pub const METRIC_GRPC_LOGICAL_PLAN: &str = "grpc.logical_plan";
pub const METRIC_GRPC_ALTER: &str = "grpc.alter";
pub const METRIC_GRPC_DROP_TABLE: &str = "grpc.drop_table";
pub const METRIC_GRPC_FLUSH_TABLE: &str = "grpc.flush_table";
pub const METRIC_GRPC_COMPACT_TABLE: &str = "grpc.compact_table";
pub const METRIC_GRPC_TRUNCATE_TABLE: &str = "grpc.truncate_table";
pub const METRIC_GRPC_DO_GET: &str = "grpc.do_get";
pub(crate) const METRIC_REGION_REQUEST_GRPC: &str = "grpc.region_request";

264
src/client/src/region.rs Normal file
View File

@@ -0,0 +1,264 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use api::v1::region::{QueryRequest, RegionRequest, RegionResponse};
use api::v1::ResponseHeader;
use arrow_flight::Ticket;
use async_stream::stream;
use async_trait::async_trait;
use common_error::ext::{BoxedError, ErrorExt};
use common_error::status_code::StatusCode;
use common_grpc::flight::{FlightDecoder, FlightMessage};
use common_meta::datanode_manager::{AffectedRows, Datanode};
use common_meta::error::{self as meta_error, Result as MetaResult};
use common_recordbatch::error::ExternalSnafu;
use common_recordbatch::{RecordBatchStreamAdaptor, SendableRecordBatchStream};
use common_telemetry::{error, timer};
use prost::Message;
use snafu::{location, Location, OptionExt, ResultExt};
use tokio_stream::StreamExt;
use crate::error::Error::RegionServer;
use crate::error::{
self, ConvertFlightDataSnafu, IllegalDatabaseResponseSnafu, IllegalFlightMessagesSnafu,
MissingFieldSnafu, Result, ServerSnafu,
};
use crate::{metrics, Client, Error};
#[derive(Debug)]
pub struct RegionRequester {
client: Client,
}
#[async_trait]
impl Datanode for RegionRequester {
async fn handle(&self, request: RegionRequest) -> MetaResult<AffectedRows> {
self.handle_inner(request).await.map_err(|err| {
if matches!(err, RegionServer { .. }) {
meta_error::Error::RetryLater {
source: BoxedError::new(err),
}
} else {
meta_error::Error::External {
source: BoxedError::new(err),
location: location!(),
}
}
})
}
async fn handle_query(&self, request: QueryRequest) -> MetaResult<SendableRecordBatchStream> {
let ticket = Ticket {
ticket: request.encode_to_vec().into(),
};
self.do_get_inner(ticket)
.await
.map_err(BoxedError::new)
.context(meta_error::ExternalSnafu)
}
}
impl RegionRequester {
pub fn new(client: Client) -> Self {
Self { client }
}
pub async fn do_get_inner(&self, ticket: Ticket) -> Result<SendableRecordBatchStream> {
let mut flight_client = self.client.make_flight_client()?;
let response = flight_client
.mut_inner()
.do_get(ticket)
.await
.map_err(|e| {
let tonic_code = e.code();
let e: error::Error = e.into();
let code = e.status_code();
let msg = e.to_string();
let error = Error::FlightGet {
tonic_code,
addr: flight_client.addr().to_string(),
source: BoxedError::new(ServerSnafu { code, msg }.build()),
};
error!(
e; "Failed to do Flight get, addr: {}, code: {}",
flight_client.addr(),
tonic_code
);
error
})?;
let flight_data_stream = response.into_inner();
let mut decoder = FlightDecoder::default();
let mut flight_message_stream = flight_data_stream.map(move |flight_data| {
flight_data
.map_err(Error::from)
.and_then(|data| decoder.try_decode(data).context(ConvertFlightDataSnafu))
});
let Some(first_flight_message) = flight_message_stream.next().await else {
return IllegalFlightMessagesSnafu {
reason: "Expect the response not to be empty",
}
.fail();
};
let FlightMessage::Schema(schema) = first_flight_message? else {
return IllegalFlightMessagesSnafu {
reason: "Expect schema to be the first flight message",
}
.fail();
};
let stream = Box::pin(stream!({
while let Some(flight_message) = flight_message_stream.next().await {
let flight_message = flight_message
.map_err(BoxedError::new)
.context(ExternalSnafu)?;
let FlightMessage::Recordbatch(record_batch) = flight_message else {
yield IllegalFlightMessagesSnafu {
reason: "A Schema message must be succeeded exclusively by a set of RecordBatch messages"
}
.fail()
.map_err(BoxedError::new)
.context(ExternalSnafu);
break;
};
yield Ok(record_batch);
}
}));
let record_batch_stream = RecordBatchStreamAdaptor {
schema,
stream,
output_ordering: None,
};
Ok(Box::pin(record_batch_stream))
}
async fn handle_inner(&self, request: RegionRequest) -> Result<AffectedRows> {
let request_type = request
.body
.as_ref()
.with_context(|| MissingFieldSnafu { field: "body" })?
.as_ref()
.to_string();
let _timer = timer!(
metrics::METRIC_REGION_REQUEST_GRPC,
&[("request_type", request_type)]
);
let mut client = self.client.raw_region_client()?;
let RegionResponse {
header,
affected_rows,
} = client
.handle(request)
.await
.map_err(|e| {
let code = e.code();
let err: error::Error = e.into();
// Uses `Error::RegionServer` instead of `Error::Server`
error::Error::RegionServer {
code,
source: BoxedError::new(err),
}
})?
.into_inner();
check_response_header(header)?;
Ok(affected_rows)
}
pub async fn handle(&self, request: RegionRequest) -> Result<AffectedRows> {
self.handle_inner(request).await
}
}
pub fn check_response_header(header: Option<ResponseHeader>) -> Result<()> {
let status = header
.and_then(|header| header.status)
.context(IllegalDatabaseResponseSnafu {
err_msg: "either response header or status is missing",
})?;
if StatusCode::is_success(status.status_code) {
Ok(())
} else {
let code =
StatusCode::from_u32(status.status_code).context(IllegalDatabaseResponseSnafu {
err_msg: format!("unknown server status: {:?}", status),
})?;
ServerSnafu {
code,
msg: status.err_msg,
}
.fail()
}
}
#[cfg(test)]
mod test {
use api::v1::Status as PbStatus;
use super::*;
use crate::Error::{IllegalDatabaseResponse, Server};
#[test]
fn test_check_response_header() {
let result = check_response_header(None);
assert!(matches!(
result.unwrap_err(),
IllegalDatabaseResponse { .. }
));
let result = check_response_header(Some(ResponseHeader { status: None }));
assert!(matches!(
result.unwrap_err(),
IllegalDatabaseResponse { .. }
));
let result = check_response_header(Some(ResponseHeader {
status: Some(PbStatus {
status_code: StatusCode::Success as u32,
err_msg: "".to_string(),
}),
}));
assert!(result.is_ok());
let result = check_response_header(Some(ResponseHeader {
status: Some(PbStatus {
status_code: u32::MAX,
err_msg: "".to_string(),
}),
}));
assert!(matches!(
result.unwrap_err(),
IllegalDatabaseResponse { .. }
));
let result = check_response_header(Some(ResponseHeader {
status: Some(PbStatus {
status_code: StatusCode::Internal as u32,
err_msg: "blabla".to_string(),
}),
}));
let Server { code, msg } = result.unwrap_err() else {
unreachable!()
};
assert_eq!(code, StatusCode::Internal);
assert_eq!(msg, "blabla");
}
}

View File

@@ -16,6 +16,7 @@ use api::v1::greptime_database_client::GreptimeDatabaseClient;
use api::v1::greptime_request::Request;
use api::v1::{
AuthHeader, GreptimeRequest, GreptimeResponse, InsertRequest, InsertRequests, RequestHeader,
RowInsertRequest, RowInsertRequests,
};
use tokio::sync::mpsc;
use tokio::task::JoinHandle;
@@ -84,6 +85,18 @@ impl StreamInserter {
})
}
pub async fn row_insert(&self, requests: Vec<RowInsertRequest>) -> Result<()> {
let inserts = RowInsertRequests { inserts: requests };
let request = self.to_rpc_request(Request::RowInserts(inserts));
self.sender.send(request).await.map_err(|e| {
error::ClientStreamingSnafu {
err_msg: e.to_string(),
}
.build()
})
}
pub async fn finish(self) -> Result<u32> {
drop(self.sender);

View File

@@ -17,13 +17,18 @@ metrics-process = ["servers/metrics-process"]
[dependencies]
anymap = "1.0.0-beta.2"
async-trait.workspace = true
auth.workspace = true
catalog = { workspace = true }
chrono.workspace = true
clap = { version = "3.1", features = ["derive"] }
client = { workspace = true }
common-base = { workspace = true }
common-catalog = { workspace = true }
common-config = { workspace = true }
common-error = { workspace = true }
common-macro = { workspace = true }
common-meta = { workspace = true }
common-procedure = { workspace = true }
common-query = { workspace = true }
common-recordbatch = { workspace = true }
common-telemetry = { workspace = true, features = [
@@ -34,17 +39,24 @@ datanode = { workspace = true }
datatypes = { workspace = true }
either = "1.8"
etcd-client.workspace = true
file-engine = { workspace = true }
frontend = { workspace = true }
futures.workspace = true
lazy_static.workspace = true
meta-client = { workspace = true }
meta-srv = { workspace = true }
metrics.workspace = true
mito2 = { workspace = true }
nu-ansi-term = "0.46"
partition = { workspace = true }
plugins.workspace = true
prost.workspace = true
query = { workspace = true }
rand.workspace = true
regex.workspace = true
rustyline = "10.1"
serde.workspace = true
serde_json.workspace = true
servers = { workspace = true }
session = { workspace = true }
snafu.workspace = true

View File

@@ -116,7 +116,7 @@ impl SubCommand {
Ok(Application::Metasrv(app))
}
(SubCommand::Standalone(cmd), Options::Standalone(opts)) => {
let app = cmd.build(opts.fe_opts, opts.dn_opts).await?;
let app = cmd.build(*opts).await?;
Ok(Application::Standalone(app))
}
(SubCommand::Cli(cmd), Options::Cli(_)) => {

View File

@@ -16,6 +16,8 @@ mod bench;
mod cmd;
mod helper;
mod repl;
// TODO(weny): Removes it
#[allow(deprecated)]
mod upgrade;
use async_trait::async_trait;
@@ -76,7 +78,7 @@ impl Command {
#[derive(Parser)]
enum SubCommand {
Attach(AttachCommand),
// Attach(AttachCommand),
Upgrade(UpgradeCommand),
Bench(BenchTableMetadataCommand),
}
@@ -84,7 +86,7 @@ enum SubCommand {
impl SubCommand {
async fn build(self) -> Result<Instance> {
match self {
SubCommand::Attach(cmd) => cmd.build().await,
// SubCommand::Attach(cmd) => cmd.build().await,
SubCommand::Upgrade(cmd) => cmd.build().await,
SubCommand::Bench(cmd) => cmd.build().await,
}
@@ -102,51 +104,9 @@ pub(crate) struct AttachCommand {
}
impl AttachCommand {
#[allow(dead_code)]
async fn build(self) -> Result<Instance> {
let repl = Repl::try_new(&self).await?;
Ok(Instance::Repl(repl))
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_load_options() {
let cmd = Command {
cmd: SubCommand::Attach(AttachCommand {
grpc_addr: String::from(""),
meta_addr: None,
disable_helper: false,
}),
};
let opts = cmd.load_options(TopLevelOptions::default()).unwrap();
let logging_opts = opts.logging_options();
assert_eq!("/tmp/greptimedb/logs", logging_opts.dir);
assert!(logging_opts.level.is_none());
assert!(!logging_opts.enable_jaeger_tracing);
}
#[test]
fn test_top_level_options() {
let cmd = Command {
cmd: SubCommand::Attach(AttachCommand {
grpc_addr: String::from(""),
meta_addr: None,
disable_helper: false,
}),
};
let opts = cmd
.load_options(TopLevelOptions {
log_dir: Some("/tmp/greptimedb/test/logs".to_string()),
log_level: Some("debug".to_string()),
})
.unwrap();
let logging_opts = opts.logging_options();
assert_eq!("/tmp/greptimedb/test/logs", logging_opts.dir);
assert_eq!("debug", logging_opts.level.as_ref().unwrap());
}
}

View File

@@ -12,53 +12,30 @@
// See the License for the specific language governing permissions and
// limitations under the License.
mod datanode_table;
mod table_info;
mod table_name;
mod table_region;
use std::collections::BTreeMap;
use std::future::Future;
use std::sync::Arc;
use std::time::{Duration, Instant};
use std::time::Duration;
use async_trait::async_trait;
use clap::Parser;
use common_meta::key::table_region::RegionDistribution;
use common_meta::key::{TableMetadataManager, TableMetadataManagerRef};
use common_meta::peer::Peer;
use common_meta::rpc::router::{Region, RegionRoute};
use common_meta::table_name::TableName;
use common_telemetry::info;
use datatypes::data_type::ConcreteDataType;
use datatypes::schema::{ColumnSchema, RawSchema};
use meta_srv::service::store::etcd::EtcdStore;
use meta_srv::service::store::kv::KvBackendAdapter;
use rand::prelude::SliceRandom;
use rand::Rng;
use table::metadata::{RawTableInfo, RawTableMeta, TableId, TableIdent, TableType};
use crate::cli::bench::datanode_table::DatanodeTableBencher;
use crate::cli::bench::table_info::TableInfoBencher;
use crate::cli::bench::table_name::TableNameBencher;
use crate::cli::bench::table_region::TableRegionBencher;
use self::metadata::TableMetadataBencher;
use crate::cli::{Instance, Tool};
use crate::error::Result;
async fn bench<F, Fut>(desc: &str, f: F, count: u32)
where
F: Fn(u32) -> Fut,
Fut: Future<Output = ()>,
{
let mut total = Duration::default();
for i in 1..=count {
let start = Instant::now();
f(i).await;
total += start.elapsed();
}
let cost = total.as_millis() as f64 / count as f64;
info!("{desc}, average operation cost: {cost:.2} ms");
}
mod metadata;
async fn bench_self_recorded<F, Fut>(desc: &str, f: F, count: u32)
where
@@ -107,31 +84,11 @@ struct BenchTableMetadata {
#[async_trait]
impl Tool for BenchTableMetadata {
async fn do_work(&self) -> Result<()> {
info!("Start benching table name manager ...");
TableNameBencher::new(self.table_metadata_manager.table_name_manager(), self.count)
.start()
.await;
info!("Start benching table info manager ...");
TableInfoBencher::new(self.table_metadata_manager.table_info_manager(), self.count)
.start()
.await;
info!("Start benching table region manager ...");
TableRegionBencher::new(
self.table_metadata_manager.table_region_manager(),
self.count,
)
.start()
.await;
info!("Start benching datanode table manager ...");
DatanodeTableBencher::new(
self.table_metadata_manager.datanode_table_manager(),
self.count,
)
.start()
.await;
let bencher = TableMetadataBencher::new(self.table_metadata_manager.clone(), self.count);
bencher.bench_create().await;
bencher.bench_get().await;
bencher.bench_rename().await;
bencher.bench_delete().await;
Ok(())
}
}
@@ -163,7 +120,6 @@ fn create_table_info(table_id: TableId, table_name: TableName) -> RawTableInfo {
created_on: chrono::DateTime::default(),
primary_key_indices: vec![],
next_column_id: columns as u32 + 1,
engine_options: Default::default(),
value_indices: vec![],
options: Default::default(),
region_numbers: (1..=100).collect(),
@@ -184,16 +140,25 @@ fn create_table_info(table_id: TableId, table_name: TableName) -> RawTableInfo {
}
}
fn create_region_distribution() -> RegionDistribution {
let mut regions = (1..=100).collect::<Vec<u32>>();
regions.shuffle(&mut rand::thread_rng());
fn create_region_routes() -> Vec<RegionRoute> {
let mut regions = Vec::with_capacity(100);
let mut rng = rand::thread_rng();
let mut region_distribution = RegionDistribution::new();
for datanode_id in 0..10 {
region_distribution.insert(
datanode_id as u64,
regions[datanode_id * 10..(datanode_id + 1) * 10].to_vec(),
);
for region_id in 0..64u64 {
regions.push(RegionRoute {
region: Region {
id: region_id.into(),
name: String::new(),
partition: None,
attrs: BTreeMap::new(),
},
leader_peer: Some(Peer {
id: rng.gen_range(0..10),
addr: String::new(),
}),
follower_peers: vec![],
});
}
region_distribution
regions
}

View File

@@ -1,131 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use common_meta::key::datanode_table::{DatanodeTableKey, DatanodeTableManager};
use super::bench;
pub struct DatanodeTableBencher<'a> {
datanode_table_manager: &'a DatanodeTableManager,
count: u32,
}
impl<'a> DatanodeTableBencher<'a> {
pub fn new(datanode_table_manager: &'a DatanodeTableManager, count: u32) -> Self {
Self {
datanode_table_manager,
count,
}
}
pub async fn start(&self) {
self.bench_create().await;
self.bench_get().await;
self.bench_move_region().await;
self.bench_tables().await;
self.bench_remove().await;
}
async fn bench_create(&self) {
let desc = format!(
"DatanodeTableBencher: create {} datanode table keys",
self.count
);
bench(
&desc,
|i| async move {
self.datanode_table_manager
.create(1, i, vec![1, 2, 3, 4])
.await
.unwrap();
},
self.count,
)
.await;
}
async fn bench_get(&self) {
let desc = format!(
"DatanodeTableBencher: get {} datanode table keys",
self.count
);
bench(
&desc,
|i| async move {
let key = DatanodeTableKey::new(1, i);
assert!(self
.datanode_table_manager
.get(&key)
.await
.unwrap()
.is_some());
},
self.count,
)
.await;
}
async fn bench_move_region(&self) {
let desc = format!(
"DatanodeTableBencher: move {} datanode table regions",
self.count
);
bench(
&desc,
|i| async move {
self.datanode_table_manager
.move_region(1, 2, i, 1)
.await
.unwrap();
},
self.count,
)
.await;
}
async fn bench_tables(&self) {
let desc = format!(
"DatanodeTableBencher: list {} datanode table keys",
self.count
);
bench(
&desc,
|_| async move {
assert!(!self
.datanode_table_manager
.tables(1)
.await
.unwrap()
.is_empty());
},
self.count,
)
.await;
}
async fn bench_remove(&self) {
let desc = format!(
"DatanodeTableBencher: remove {} datanode table keys",
self.count
);
bench(
&desc,
|i| async move {
self.datanode_table_manager.remove(1, i).await.unwrap();
},
self.count,
)
.await;
}
}

View File

@@ -0,0 +1,136 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::time::Instant;
use common_meta::key::TableMetadataManagerRef;
use common_meta::table_name::TableName;
use super::{bench_self_recorded, create_region_routes, create_table_info};
pub struct TableMetadataBencher {
table_metadata_manager: TableMetadataManagerRef,
count: u32,
}
impl TableMetadataBencher {
pub fn new(table_metadata_manager: TableMetadataManagerRef, count: u32) -> Self {
Self {
table_metadata_manager,
count,
}
}
pub async fn bench_create(&self) {
let desc = format!(
"TableMetadataBencher: creating {} table metadata",
self.count
);
bench_self_recorded(
&desc,
|i| async move {
let table_name = format!("bench_table_name_{}", i);
let table_name = TableName::new("bench_catalog", "bench_schema", table_name);
let table_info = create_table_info(i, table_name);
let region_routes = create_region_routes();
let start = Instant::now();
self.table_metadata_manager
.create_table_metadata(table_info, region_routes)
.await
.unwrap();
start.elapsed()
},
self.count,
)
.await;
}
pub async fn bench_get(&self) {
let desc = format!(
"TableMetadataBencher: getting {} table info and region routes",
self.count
);
bench_self_recorded(
&desc,
|i| async move {
let start = Instant::now();
self.table_metadata_manager
.get_full_table_info(i)
.await
.unwrap();
start.elapsed()
},
self.count,
)
.await;
}
pub async fn bench_delete(&self) {
let desc = format!(
"TableMetadataBencher: deleting {} table metadata",
self.count
);
bench_self_recorded(
&desc,
|i| async move {
let (table_info, table_route) = self
.table_metadata_manager
.get_full_table_info(i)
.await
.unwrap();
let start = Instant::now();
let _ = self
.table_metadata_manager
.delete_table_metadata(&table_info.unwrap(), &table_route.unwrap())
.await;
start.elapsed()
},
self.count,
)
.await;
}
pub async fn bench_rename(&self) {
let desc = format!("TableMetadataBencher: renaming {} table", self.count);
bench_self_recorded(
&desc,
|i| async move {
let (table_info, _) = self
.table_metadata_manager
.get_full_table_info(i)
.await
.unwrap();
let new_table_name = format!("renamed_{}", i);
let start = Instant::now();
let _ = self
.table_metadata_manager
.rename_table(table_info.unwrap(), new_table_name)
.await;
start.elapsed()
},
self.count,
)
.await;
}
}

View File

@@ -1,111 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::time::Instant;
use common_meta::key::table_info::TableInfoManager;
use common_meta::table_name::TableName;
use super::{bench, bench_self_recorded, create_table_info};
pub struct TableInfoBencher<'a> {
table_info_manager: &'a TableInfoManager,
count: u32,
}
impl<'a> TableInfoBencher<'a> {
pub fn new(table_info_manager: &'a TableInfoManager, count: u32) -> Self {
Self {
table_info_manager,
count,
}
}
pub async fn start(&self) {
self.bench_create().await;
self.bench_get().await;
self.bench_compare_and_put().await;
self.bench_remove().await;
}
async fn bench_create(&self) {
let desc = format!("TableInfoBencher: create {} table infos", self.count);
bench(
&desc,
|i| async move {
let table_name = format!("bench_table_name_{}", i);
let table_name = TableName::new("bench_catalog", "bench_schema", table_name);
let table_info = create_table_info(i, table_name);
self.table_info_manager
.create(i, &table_info)
.await
.unwrap();
},
self.count,
)
.await;
}
async fn bench_get(&self) {
let desc = format!("TableInfoBencher: get {} table infos", self.count);
bench(
&desc,
|i| async move {
assert!(self.table_info_manager.get(i).await.unwrap().is_some());
},
self.count,
)
.await;
}
async fn bench_compare_and_put(&self) {
let desc = format!(
"TableInfoBencher: compare_and_put {} table infos",
self.count
);
bench_self_recorded(
&desc,
|i| async move {
let table_info_value = self.table_info_manager.get(i).await.unwrap().unwrap();
let mut new_table_info = table_info_value.table_info.clone();
new_table_info.ident.version += 1;
let start = Instant::now();
self.table_info_manager
.compare_and_put(i, Some(table_info_value), new_table_info)
.await
.unwrap()
.unwrap();
start.elapsed()
},
self.count,
)
.await;
}
async fn bench_remove(&self) {
let desc = format!("TableInfoBencher: remove {} table infos", self.count);
bench(
&desc,
|i| async move {
self.table_info_manager.remove(i).await.unwrap();
},
self.count,
)
.await;
}
}

View File

@@ -1,131 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use common_meta::key::table_name::{TableNameKey, TableNameManager};
use super::bench;
pub struct TableNameBencher<'a> {
table_name_manager: &'a TableNameManager,
count: u32,
}
impl<'a> TableNameBencher<'a> {
pub fn new(table_name_manager: &'a TableNameManager, count: u32) -> Self {
Self {
table_name_manager,
count,
}
}
pub async fn start(&self) {
self.bench_create().await;
self.bench_rename().await;
self.bench_get().await;
self.bench_tables().await;
self.bench_remove().await;
}
async fn bench_create(&self) {
let desc = format!("TableNameBencher: create {} table names", self.count);
bench(
&desc,
|i| async move {
let table_name = format!("bench_table_name_{}", i);
let table_name_key = create_table_name_key(&table_name);
self.table_name_manager
.create(&table_name_key, i)
.await
.unwrap();
},
self.count,
)
.await;
}
async fn bench_rename(&self) {
let desc = format!("TableNameBencher: rename {} table names", self.count);
bench(
&desc,
|i| async move {
let table_name = format!("bench_table_name_{}", i);
let new_table_name = format!("bench_table_name_new_{}", i);
let table_name_key = create_table_name_key(&table_name);
self.table_name_manager
.rename(table_name_key, i, &new_table_name)
.await
.unwrap();
},
self.count,
)
.await;
}
async fn bench_get(&self) {
let desc = format!("TableNameBencher: get {} table names", self.count);
bench(
&desc,
|i| async move {
let table_name = format!("bench_table_name_new_{}", i);
let table_name_key = create_table_name_key(&table_name);
assert!(self
.table_name_manager
.get(table_name_key)
.await
.unwrap()
.is_some());
},
self.count,
)
.await;
}
async fn bench_tables(&self) {
let desc = format!("TableNameBencher: list all {} table names", self.count);
bench(
&desc,
|_| async move {
assert!(!self
.table_name_manager
.tables("bench_catalog", "bench_schema")
.await
.unwrap()
.is_empty());
},
self.count,
)
.await;
}
async fn bench_remove(&self) {
let desc = format!("TableNameBencher: remove {} table names", self.count);
bench(
&desc,
|i| async move {
let table_name = format!("bench_table_name_new_{}", i);
let table_name_key = create_table_name_key(&table_name);
self.table_name_manager
.remove(table_name_key)
.await
.unwrap();
},
self.count,
)
.await;
}
}
fn create_table_name_key(table_name: &str) -> TableNameKey {
TableNameKey::new("bench_catalog", "bench_schema", table_name)
}

View File

@@ -1,112 +0,0 @@
// Copyright 2023 Greptime Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
use std::time::Instant;
use common_meta::key::table_region::TableRegionManager;
use super::{bench, bench_self_recorded, create_region_distribution};
pub struct TableRegionBencher<'a> {
table_region_manager: &'a TableRegionManager,
count: u32,
}
impl<'a> TableRegionBencher<'a> {
pub fn new(table_region_manager: &'a TableRegionManager, count: u32) -> Self {
Self {
table_region_manager,
count,
}
}
pub async fn start(&self) {
self.bench_create().await;
self.bench_get().await;
self.bench_compare_and_put().await;
self.bench_remove().await;
}
async fn bench_create(&self) {
let desc = format!("TableRegionBencher: create {} table regions", self.count);
bench_self_recorded(
&desc,
|i| async move {
let region_distribution = create_region_distribution();
let start = Instant::now();
self.table_region_manager
.create(i, &region_distribution)
.await
.unwrap();
start.elapsed()
},
self.count,
)
.await;
}
async fn bench_get(&self) {
let desc = format!("TableRegionBencher: get {} table regions", self.count);
bench(
&desc,
|i| async move {
assert!(self.table_region_manager.get(i).await.unwrap().is_some());
},
self.count,
)
.await;
}
async fn bench_compare_and_put(&self) {
let desc = format!(
"TableRegionBencher: compare_and_put {} table regions",
self.count
);
bench_self_recorded(
&desc,
|i| async move {
let table_region_value = self.table_region_manager.get(i).await.unwrap().unwrap();
let new_region_distribution = create_region_distribution();
let start = Instant::now();
self.table_region_manager
.compare_and_put(i, Some(table_region_value), new_region_distribution)
.await
.unwrap()
.unwrap();
start.elapsed()
},
self.count,
)
.await;
}
async fn bench_remove(&self) {
let desc = format!("TableRegionBencher: remove {} table regions", self.count);
bench(
&desc,
|i| async move {
assert!(self.table_region_manager.remove(i).await.unwrap().is_some());
},
self.count,
)
.await;
}
}

View File

@@ -16,20 +16,16 @@ use std::path::PathBuf;
use std::sync::Arc;
use std::time::Instant;
use catalog::remote::CachedMetaKvBackend;
use catalog::kvbackend::{CachedMetaKvBackend, KvBackendCatalogManager};
use client::client_manager::DatanodeClients;
use client::{Client, Database, DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
use common_base::Plugins;
use common_error::ext::ErrorExt;
use common_meta::key::TableMetadataManager;
use common_query::Output;
use common_recordbatch::RecordBatches;
use common_telemetry::logging;
use either::Either;
use frontend::catalog::FrontendCatalogManager;
use meta_client::client::MetaClientBuilder;
use partition::manager::PartitionRuleManager;
use partition::route::TableRoutes;
use query::datafusion::DatafusionQueryEngine;
use query::logical_optimizer::LogicalOptimizer;
use query::parser::QueryLanguageParser;
@@ -39,7 +35,7 @@ use query::QueryEngine;
use rustyline::error::ReadlineError;
use rustyline::Editor;
use session::context::QueryContext;
use snafu::{ErrorCompat, ResultExt};
use snafu::ResultExt;
use substrait::{DFLogicalSubstraitConvertor, SubstraitPlan};
use crate::cli::cmd::ReplCommand;
@@ -152,7 +148,7 @@ impl Repl {
.await
.map_err(|e| {
let status_code = e.status_code();
let root_cause = e.iter_chain().last().unwrap();
let root_cause = e.output_msg();
println!("Error: {}({status_code}), {root_cause}", status_code as u32)
})
.is_ok()
@@ -180,7 +176,7 @@ impl Repl {
.encode(&plan)
.context(SubstraitEncodeLogicalPlanSnafu)?;
self.database.logical_plan(plan.to_vec(), None).await
self.database.logical_plan(plan.to_vec(), 0).await
} else {
self.database.sql(&sql).await
}
@@ -254,24 +250,19 @@ async fn create_query_engine(meta_addr: &str) -> Result<DatafusionQueryEngine> {
let cached_meta_backend = Arc::new(CachedMetaKvBackend::new(meta_client.clone()));
let table_routes = Arc::new(TableRoutes::new(meta_client));
let partition_manager = Arc::new(PartitionRuleManager::new(table_routes));
let datanode_clients = Arc::new(DatanodeClients::default());
let catalog_list = Arc::new(FrontendCatalogManager::new(
let catalog_list = KvBackendCatalogManager::new(
cached_meta_backend.clone(),
cached_meta_backend.clone(),
partition_manager,
datanode_clients,
Arc::new(TableMetadataManager::new(cached_meta_backend)),
));
let plugins: Arc<Plugins> = Default::default();
);
let plugins: Plugins = Default::default();
let state = Arc::new(QueryEngineState::new(
catalog_list,
None,
None,
false,
None,
None,
plugins.clone(),
));

Some files were not shown because too many files have changed in this diff Show More