Compare commits

...

77 Commits

Author SHA1 Message Date
discord9
365421c452 Revert "feat: stream drop record metrics"
This reverts commit 3eda4a2257928d95cf9c1328ae44fae84cfbb017.

Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
b361c5f50a chore: more dbg
Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
5d78bc1efa test: update sqlness
Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
99c78b2f97 fix: expand differently
Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
bdecdb869e feat: stream drop record metrics
Signed-off-by: discord9 <discord9@163.com>

refactor: move logging to drop too

Signed-off-by: discord9 <discord9@163.com>

fix: drop input stream before collect metrics

Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
df4cd157e1 Revert "feat: stream drop record metrics"
This reverts commit 6a16946a5b8ea37557bbb1b600847d24274d6500.

Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
60fbe54f90 feat: stream drop record metrics
Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
7c15e71407 revert
Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
13f20b5a40 add logging to figure test failure
Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
Ruihang Xia
2b802e45f5 update sqlness result
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
Ruihang Xia
def9b7c01d fix: expand on conditional commutative as well
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Signed-off-by: discord9 <discord9@163.com>
2025-07-10 17:23:49 +08:00
discord9
af03e89139 fix: stricter win sort condition (#6477)
test: sqlness



test: fix sqlness redacted

Signed-off-by: discord9 <discord9@163.com>
2025-07-08 22:27:17 +00:00
jeremyhi
e7a64b7dc0 chore: refactor register_region method to avoid TOCTOU issues (#6468) 2025-07-08 13:26:38 +00:00
Lin Yihai
29739b556e refactor: split some convert function into sql-common crate (#6452)
refactor: split some convert function into `sql-common` crates

Signed-off-by: Yihai Lin <yihai-lin@foxmail.com>
2025-07-08 12:08:33 +00:00
Lei, HUANG
77e50d0e08 chore: expose some config (#6479)
refactor/expose-config:
 ### Make SubCommand and Fields Public in `frontend.rs`

 - Made `subcmd` field in `Command` struct public.
 - Made `SubCommand` enum public.
 - Made `config_file` and `env_prefix` fields in `StartCommand` struct public.

 These changes enhance the accessibility of command-related structures and fields, facilitating external usage and integration.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
2025-07-08 11:52:23 +00:00
LFC
c2f1447345 refactor: stores the http server builder in Metasrv instance (#6461)
* refactor: stores the http server builder in Metasrv instance

Signed-off-by: luofucong <luofc@foxmail.com>

* resolve PR comments

Signed-off-by: luofucong <luofc@foxmail.com>

* fix ci

Signed-off-by: luofucong <luofc@foxmail.com>

---------

Signed-off-by: luofucong <luofc@foxmail.com>
2025-07-07 06:39:05 +00:00
Weny Xu
30f7955d2b feat: add column metadata to response extensions (#6451)
Signed-off-by: WenyXu <wenymedia@gmail.com>
2025-07-07 03:38:13 +00:00
fys
3508fddd74 feat: support show triggers (#6465)
* feat: support show triggers

* add enterprise feature

* chore: remove unused error
2025-07-07 03:30:57 +00:00
Weny Xu
351c741c70 fix(metric-engine): handle stale metadata region recovery failures (#6395)
* fix(metric-engine): handle stale metadata region recovery failures

Signed-off-by: WenyXu <wenymedia@gmail.com>

* test: add unit tests

Signed-off-by: WenyXu <wenymedia@gmail.com>

---------

Signed-off-by: WenyXu <wenymedia@gmail.com>
2025-07-07 03:30:40 +00:00
liyang
bb43d604a4 ci: use ubuntu 16 core machine in release-cn-artifacts (#6464)
ci: use ubuntu 16 core machine
2025-07-04 18:06:07 +00:00
Lei, HUANG
9576bcb9ae fix: filter empty batch in bulk insert api (#6459)
* fix/filter-empty-batch-in-bulk-insert-api:
 **Add Early Return for Empty Record Batches in `bulk_insert.rs`**

 - Implemented an early return in the `Inserter` implementation to handle cases where `record_batch.num_rows()` is zero, improving efficiency by avoiding unnecessary processing.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

* fix/filter-empty-batch-in-bulk-insert-api:
 **Improve Bulk Insert Handling**

 - **`handle_bulk_insert.rs`**: Added a check to handle cases where the batch has zero rows, immediately returning and sending a success response with zero rows processed.
 - **`bulk_insert.rs`**: Enhanced logic to skip processing for masks that select none, optimizing the bulk insert operation by avoiding unnecessary iterations.

 These changes improve the efficiency and robustness of the bulk insert process by handling edge cases more effectively.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

* fix/filter-empty-batch-in-bulk-insert-api:
 ### Refactor and Error Handling Enhancements

 - **Refactored Timestamp Handling**: Introduced `timestamp_array_to_primitive` function in `timestamp.rs` to streamline conversion of timestamp arrays to primitive arrays, reducing redundancy in `handle_bulk_insert.rs` and `bulk_insert.rs`.
 - **Error Handling**: Added `InconsistentTimestampLength` error in `error.rs` to handle mismatched timestamp column lengths in bulk insert operations.
 - **Bulk Insert Logic**: Updated `handle_bulk_insert.rs` to utilize the new timestamp conversion function and added checks for timestamp length consistency.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

* fix/filter-empty-batch-in-bulk-insert-api:
 **Refactor `bulk_insert.rs` to streamline imports**

 - Simplified import statements by removing unused timestamp-related arrays and data types from the `arrow` crate in `bulk_insert.rs`.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

---------

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
2025-07-04 13:32:10 +00:00
Zhenchi
dc17e6e517 fix: add backward compatibility for SkippingIndexOptions deserialization (#6458)
* fix: add backward compatibility for `SkippingIndexOptions` deserialization

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

* address comments

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

* address comments

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

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2025-07-04 11:58:27 +00:00
Yiran
563d25ee04 fix: doc links (#6305)
Signed-off-by: Yiran <cuiyiran3@gmail.com>
2025-07-04 09:52:47 +00:00
Weny Xu
7d17782fd5 feat: persist column ids in table metadata (#6457)
* feat: persist column ids in table metadata

Signed-off-by: WenyXu <wenymedia@gmail.com>

* chore: apply suggestions from CR

Signed-off-by: WenyXu <wenymedia@gmail.com>

---------

Signed-off-by: WenyXu <wenymedia@gmail.com>
2025-07-04 08:12:29 +00:00
fys
c5360601f5 feat: information table extension (#6434)
* feat: information table extension

* avoid use std HashMap behind cfg feature
2025-07-04 04:37:36 +00:00
discord9
9b5baa965c feat: truly limit time range by split window (#6295)
* feat: actually split window to limit time range

feat: truly limit time range by split window

Update src/flow/src/batching_mode/state.rs

Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>
Signed-off-by: discord9 <discord9@163.com>

* chore: added stalled time window range

Signed-off-by: discord9 <discord9@163.com>

* fix: not flush all time range as too expensive

Signed-off-by: discord9 <discord9@163.com>

* test: make it more robust

Signed-off-by: discord9 <discord9@163.com>

* what

Signed-off-by: discord9 <discord9@163.com>

* feat: denfensively handle surplus

Signed-off-by: discord9 <discord9@163.com>

* refactor: per review,explain flush flow

Signed-off-by: discord9 <discord9@163.com>

* chore: per bugbot

Signed-off-by: discord9 <discord9@163.com>

* fix: a temp fix to make mirror insert go first(still need better fix to sync with mirror insert that happens before

Signed-off-by: discord9 <discord9@163.com>

* chore: add todo

Signed-off-by: discord9 <discord9@163.com>

---------

Signed-off-by: discord9 <discord9@163.com>
Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>
2025-07-04 03:37:43 +00:00
Yingwen
76a5145def fix: enable max_execution time for other read only statements (#6454)
Also disable the timeout when timeout is 0

Signed-off-by: evenyag <realevenyag@gmail.com>
2025-07-03 13:46:02 +00:00
Ruihang Xia
7b2703760b feat: skip rule checker on ingestion (#6453)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2025-07-03 13:31:16 +00:00
Ruihang Xia
81ea172ce4 feat!: point matrix based partition rule checker (#6431)
* bare implementation

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

* stateful generator

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

* error report

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

* fix remap checkpoint

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

* use matrix generator as iterator

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

* pre-calculate suffix product

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

* update existing test cases

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

* fix clippy

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

* sqlness

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

* fix ut

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

* clean up

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

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2025-07-03 06:50:02 +00:00
dennis zhuang
f7c363f969 fix: label_replace and label_join functions when used as sub‐expressions (#6443)
* fix: label_replace and label_join functions in expressions

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* chore: remove update_fields

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* chore: tql eval -> TQL EVAL

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* fix: empty regex and not existing source label

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* chore: simplfy test

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* fix: test

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* fix: test

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

---------

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
2025-07-03 05:34:22 +00:00
Ruihang Xia
5f2daae087 fix: remap column indices on overriding logical table partitions (#6446)
* fix: remap column indices on overriding logical table partitions

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

* sqlness

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

* refactor map query

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

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2025-07-02 12:12:00 +00:00
Yingwen
b1b0d0136f fix: correct MAX_EXECUTION_TIME timeout calculation (#6444)
* feat: implement statement timeout in frontend instance

Signed-off-by: evenyag <realevenyag@gmail.com>

* fix: fail fast when timeout is 0

Signed-off-by: evenyag <realevenyag@gmail.com>

* fix: update start time

Signed-off-by: evenyag <realevenyag@gmail.com>

---------

Signed-off-by: evenyag <realevenyag@gmail.com>
2025-07-02 08:31:40 +00:00
Zhenchi
599f289f59 feat: add granularity and false_positive_rate options for indexes (#6416)
* feat: add `granularity` and `false_positive_rate` options for indexes

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

* address comments

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

* upgrade proto

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

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2025-07-02 07:33:39 +00:00
LFC
385f12a62e refactor: extract the common method for errors into tonic status (#6437)
Signed-off-by: luofucong <luofc@foxmail.com>
2025-07-02 02:57:30 +00:00
fys
6b90e2b6b4 fix: allow clippy::print_stdout in cli crate (#6436)
* fix: allow clippy::print_stdout in cli crate

* add clippy lint options
2025-07-02 01:40:58 +00:00
ZonaHe
a4f3e96e96 feat: update dashboard to v0.10.2 (#6433)
Co-authored-by: sunchanglong <sunchanglong@users.noreply.github.com>
2025-07-02 01:27:37 +00:00
Ruihang Xia
2b0f27da51 feat: don't allow creating flow with the same sink and source table (#6435)
Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2025-07-01 11:33:09 +00:00
Weny Xu
e0382eeb7c fix: fix dest_keys chunks bug in TombstoneManager (#6432)
* fix(meta): fix dest_keys_chunks bug in TombstoneManager

Signed-off-by: WenyXu <wenymedia@gmail.com>

* chore: fix typo

Signed-off-by: WenyXu <wenymedia@gmail.com>

* fix: fix sqlness tests

Signed-off-by: WenyXu <wenymedia@gmail.com>

---------

Signed-off-by: WenyXu <wenymedia@gmail.com>
2025-07-01 09:20:13 +00:00
liyang
4aa6add8dc ci: add check-version script to check whether push the latast image (#6415)
Signed-off-by: liyang <daviderli614@gmail.com>
2025-07-01 07:45:47 +00:00
zyy17
645988975e refactor: add RegionMigrationTriggerReason in RegionMigrationProcedureTask (#6413)
Signed-off-by: zyy17 <zyylsxm@gmail.com>
2025-06-30 11:16:41 +00:00
LFC
a203909de3 feat: extension range definition (#6386)
* feat: defined extension range

Signed-off-by: luofucong <luofc@foxmail.com>

* remove feature parameters

Signed-off-by: luofucong <luofc@foxmail.com>

* resolve PR comments

Signed-off-by: luofucong <luofc@foxmail.com>

* resolve PR comments

Signed-off-by: luofucong <luofc@foxmail.com>

---------

Signed-off-by: luofucong <luofc@foxmail.com>
2025-06-30 02:42:40 +00:00
discord9
616e76941a feat: flow query parallel=1&query faster with many windows&min one time window (#6324)
* feat: flow query parallel=1&query faster when
windows too many&min one time window

Signed-off-by: discord9 <discord9@163.com>

* chore: default flow query parallelism=1

Signed-off-by: discord9 <discord9@163.com>

* refactor: use query options in flownode per review

Signed-off-by: discord9 <discord9@163.com>

* docs: update comment

Signed-off-by: discord9 <discord9@163.com>

* chore: fix test

Signed-off-by: discord9 <discord9@163.com>

* chore: per review

Signed-off-by: discord9 <discord9@163.com>

* chore: make config docs

Signed-off-by: discord9 <discord9@163.com>

---------

Signed-off-by: discord9 <discord9@163.com>
2025-06-30 02:17:01 +00:00
Yingwen
bc42d35c2a chore: bump version to 0.16 (#6417)
Signed-off-by: evenyag <realevenyag@gmail.com>
2025-06-28 01:46:01 +00:00
fys
524bdfff22 fix: add cfg for DecodeSqlValue error (#6420) 2025-06-28 01:39:06 +00:00
fys
6bed0b6ba0 feat: add trigger-related error code (#6419) 2025-06-28 01:25:20 +00:00
shuiyisong
dec8c52b18 feat(pipeline): support Loki API (#6390)
* chore: use schema_info

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* refactor: abstract loki item generator

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* chore: introduce middle item

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* feat: introduce pipeline in loki api

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* test: add tests

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* chore: minor update

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* chore: minor update

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* chore: update prefix and test

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* chore: change recursion to loop

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

* fix: cr issue

Signed-off-by: shuiyisong <xixing.sys@gmail.com>

---------

Signed-off-by: shuiyisong <xixing.sys@gmail.com>
2025-06-28 01:01:08 +00:00
zyy17
753a7e1a24 refactor: pass pipeline name through http header and get db from query context (#6405)
Signed-off-by: zyy17 <zyylsxm@gmail.com>
2025-06-27 10:43:37 +00:00
Lei, HUANG
6684200fce fix: skip failing nodes when gathering porcess info (#6412)
* fix/process-manager-skip-fail-nodes:
 - **Enhance Error Handling in `process_manager.rs`:**
   Improved error handling by adding a warning log for failing nodes in the `list_process` method. This ensures that the process listing continues even if some nodes fail to respond.

 - **Add Error Type Import in `process_manager.rs`:**
   Included the `Error` type from the `error` module to handle errors more effectively within the `ProcessManager` implementation.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

* fix: clippy

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

* fix/process-manager-skip-fail-nodes:
 **Enhancements to Debugging and Trait Implementation**

 - **`process_manager.rs`**: Improved logging by adding more detailed error messages when skipping failing nodes.
 - **`selector.rs`**: Enhanced the `FrontendClient` trait by adding the `Debug` trait bound to improve debugging capabilities.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

---------

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
2025-06-27 08:20:01 +00:00
Weny Xu
5fcb97724f chore: correct typo in configuration (#6411)
Signed-off-by: WenyXu <wenymedia@gmail.com>
2025-06-27 07:55:13 +00:00
Zhenchi
ff559b2688 fix: complete partial index search results in cache (#6403)
* fix: complete partial index search results in cache

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

* polish

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

* address comments

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

* add initial tests

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

* cover issue case

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

* TestEnv new -> async

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

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
2025-06-27 07:40:14 +00:00
Ruihang Xia
8473a34fc9 feat: Collider for playing with PartitionRule (#6399)
* skeleton

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

* initial impl and tests

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

* refactor and reorganize

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

* fix clippy

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

* fix typo

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

* add comment

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

* error handling

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

* explain naming

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

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2025-06-27 07:15:33 +00:00
jeremyhi
df0ebf0378 feat: override logical table's partition key indices (#6385)
* feat: Override logical table's partition key indices with physical table's

* chore: by comment
2025-06-27 02:55:56 +00:00
Weny Xu
4a665fd27b refactor: move #[allow(clippy::print_stdout)] to lib level (#6398)
chore: allow cli to print stdout

Signed-off-by: WenyXu <wenymedia@gmail.com>
2025-06-27 02:40:12 +00:00
Yingwen
b4d6441716 refactor: rename test show_processList to show_process_list (#6408)
refactor: rename show_processList to show_process_list

Signed-off-by: evenyag <realevenyag@gmail.com>
2025-06-27 01:18:54 +00:00
liyang
bdd50a2263 ci: try to fix the job permissions (#6407)
Signed-off-by: liyang <daviderli614@gmail.com>
2025-06-26 17:23:43 +00:00
codephage
f87b12b2aa feat: remove own pow fn (#6404)
feat remove own pow fn

Signed-off-by: codephage. <381510760@qq.com>
2025-06-26 09:27:30 +00:00
Yiran
07eec083b9 fix: doc issue assignee (#6406)
Signed-off-by: Yiran <cuiyiran3@gmail.com>
2025-06-26 09:18:47 +00:00
Weny Xu
4737285275 feat: implement pause/resume functionality for procedure manager (#6393)
* feat: implement pause/resume functionality for procedure manager

Signed-off-by: WenyXu <wenymedia@gmail.com>

* chore: apply suggestions from CR

Signed-off-by: WenyXu <wenymedia@gmail.com>

* chore: apply suggestions from CR

Signed-off-by: WenyXu <wenymedia@gmail.com>

---------

Signed-off-by: WenyXu <wenymedia@gmail.com>
2025-06-26 01:57:12 +00:00
codephage
55f5e09885 fix: sqlness_test show_processList (#6401)
fix test sqlness_test show_processList

Signed-off-by: codephage. <381510760@qq.com>
2025-06-26 01:53:55 +00:00
liyang
9ab36e9a6f test: add a test load configuration example for flownode (#6397)
* test: add a test load configuration example for flownode

Signed-off-by: liyang <daviderli614@gmail.com>

* format rust

Signed-off-by: liyang <daviderli614@gmail.com>

* fix cargo clippy

Signed-off-by: liyang <daviderli614@gmail.com>

* refine FlownodeOptions visibility

Signed-off-by: liyang <daviderli614@gmail.com>

* format rust

Signed-off-by: liyang <daviderli614@gmail.com>

---------

Signed-off-by: liyang <daviderli614@gmail.com>
2025-06-26 01:48:53 +00:00
Lei, HUANG
4bb5d00a4b fix(http): apply string validation mode to pipeline processor (#6378)
* fix/apply-string-validation-to-pipeline:
 ### Commit Summary

 - **Refactor `decode_string` Functionality**:
   - Moved `decode_string` logic into `PromValidationMode` as a method `decode_string`.
   - Updated all references to use the new method.
   - Files affected: `http.rs`, `prom_row_builder.rs`, `proto.rs`.

 - **Logging Enhancements**:
   - Added `debug` logging for invalid UTF-8 string values.
   - File affected: `http.rs`.

 - **Test Updates**:
   - Modified tests to use the new `decode_string` method in `PromValidationMode`.
   - File affected: `proto.rs`.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

* fix clippy

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>

---------

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
2025-06-25 18:56:35 +00:00
Lei, HUANG
1d07864b29 refactor(object-store): move backends building functions back to object-store (#6400)
refactor/building-backend-in-object-store:
 ### Refactor Object Store Configuration

 - **Centralize Object Store Configurations**: Moved object store configurations (`FileConfig`, `S3Config`, `OssConfig`, `AzblobConfig`, `GcsConfig`) to `object-store/src/config.rs`.
 - **Error Handling Enhancements**: Introduced `object-store/src/error.rs` for improved error handling related to object store operations.
 - **Factory Pattern for Object Store**: Implemented `object-store/src/factory.rs` to create object store instances, consolidating logic from `datanode/src/store.rs`.
 - **Remove Redundant Store Implementations**: Deleted individual store files (`azblob.rs`, `fs.rs`, `gcs.rs`, `oss.rs`, `s3.rs`) from `datanode/src/store/`.
 - **Update Usage of Object Store Config**: Updated references to `ObjectStoreConfig` in `datanode.rs`, `standalone.rs`, `config.rs`, and `error.rs` to use the new centralized configuration.

Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
2025-06-25 13:49:55 +00:00
ZonaHe
9be75361a4 feat: update dashboard to v0.10.1 (#6396)
Co-authored-by: ZonaHex <ZonaHex@users.noreply.github.com>
2025-06-25 12:58:04 +00:00
Ning Sun
9c1df68a5f feat: introduce /v1/health for healthcheck from external (#6388)
Signed-off-by: Ning Sun <sunning@greptime.com>
2025-06-25 12:25:36 +00:00
fys
0209461155 chore: add components for standalone instance (#6383) 2025-06-25 12:17:34 +00:00
Ruihang Xia
e728cb33fb feat: implement count_hash aggr function (#6342)
* feat: implement count_hash aggr function

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

* sqlness case

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

* change copyright year

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

* review changes

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

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
2025-06-25 12:14:03 +00:00
fys
cde7e11983 refactor: avoid adding feature to parameter (#6391)
* refactor: avoid adding feature to parameter

* avoid `cfg(not(feature = ...))` block
2025-06-25 10:47:20 +00:00
dennis zhuang
944b4b3e49 feat: supports CsvWithNames and CsvWithNamesAndTypes formats (#6384)
* feat: supports CsvWithNames and CsvWithNamesAndTypes formats and object/array types

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* test: added and fixed tests

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* chore: fix test

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* chore: remove comments

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* test: add json type csv tests

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>

* chore: remove comment

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

---------

Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
Co-authored-by: Yingwen <realevenyag@gmail.com>
2025-06-25 07:28:11 +00:00
fys
7953b090c0 feat: support syntax parsing of drop trigger (#6371)
* feat: trigger drop

* Update Cargo.toml

* Update Cargo.lock

---------

Co-authored-by: Ning Sun <classicning@gmail.com>
2025-06-24 18:48:39 +00:00
liyang
7aa9af5ba6 chore: clarify default OTLP endpoint value (#6381)
Signed-off-by: liyang <daviderli614@gmail.com>
2025-06-24 11:44:45 +00:00
Ruihang Xia
7a9444c85b refactor: remove staled manifest structures (#6382)
* refactor: remove staled manifest structures

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

* Update src/store-api/src/lib.rs

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2025-06-24 09:23:10 +00:00
LFC
bb12be3310 refactor: scan Batches directly (#6369)
* refactor: scan `Batch`es directly

Signed-off-by: luofucong <luofc@foxmail.com>

* fix ci

Signed-off-by: luofucong <luofc@foxmail.com>

* resolve PR comments

Signed-off-by: luofucong <luofc@foxmail.com>

* resolve PR comments

Signed-off-by: luofucong <luofc@foxmail.com>

---------

Signed-off-by: luofucong <luofc@foxmail.com>
2025-06-24 07:55:49 +00:00
Weny Xu
24019334ee feat: implement automatic region failure detector registrations (#6370)
* feat: implement automatic region failure detector registrations

Signed-off-by: WenyXu <wenymedia@gmail.com>

* chore: remove unused error

Signed-off-by: WenyXu <wenymedia@gmail.com>

* test: add more tests

Signed-off-by: WenyXu <wenymedia@gmail.com>

* feat: add `region_failure_detector_initialization_delay` option

Signed-off-by: WenyXu <wenymedia@gmail.com>

* chore: update config.md

Signed-off-by: WenyXu <wenymedia@gmail.com>

* chore: apply suggestions from CR

Signed-off-by: WenyXu <wenymedia@gmail.com>

* chore: update config.md

Signed-off-by: WenyXu <wenymedia@gmail.com>

---------

Signed-off-by: WenyXu <wenymedia@gmail.com>
2025-06-24 06:12:12 +00:00
codephage
116d5cf82b feat: support mysql flavor show processlist shortcut (#6328) (#6379)
* feat: support mysql flavor show processlist shortcut (#6328)

Signed-off-by: codephage. <381510760@qq.com>

* Refactor SHOW PROCESSLIST handling and add tests

Signed-off-by: codephage. <381510760@qq.com>

* add sqlness test

Signed-off-by: codephage. <381510760@qq.com>

* add sqlness test result

Signed-off-by: codephage. <381510760@qq.com>

* fix sqlness test show_processList

Signed-off-by: codephage. <381510760@qq.com>

---------

Signed-off-by: codephage. <381510760@qq.com>
2025-06-24 03:50:16 +00:00
zyy17
90a3894564 refactor: always write parent_span_id for otlp traces ingestion (#6356)
* refactor: always write `parent_span_id` for otlp traces ingestion

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

* refactor: support to write None value in row writer

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

---------

Signed-off-by: zyy17 <zyylsxm@gmail.com>
2025-06-23 11:36:57 +00:00
Yingwen
39d3e0651d feat: Support ListMetadataRequest to retrieve regions' metadata (#6348)
* feat: support list metadata in region server

Signed-off-by: evenyag <realevenyag@gmail.com>

* test: add test for list region metadata

Signed-off-by: evenyag <realevenyag@gmail.com>

* feat: return null if region not exists

Signed-off-by: evenyag <realevenyag@gmail.com>

* chore: update greptime-proto

Signed-off-by: evenyag <realevenyag@gmail.com>

---------

Signed-off-by: evenyag <realevenyag@gmail.com>
2025-06-23 07:11:20 +00:00
zyy17
a49edc6ca6 refactor: add otlp_export_protocol config to support export trace data through gRPC and HTTP protocol (#6357)
* refactor: support http traces

* refactor: add `otlp_export_protocol` config to support export trace data through gRPC and HTTP protocol

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

* Update src/common/telemetry/src/logging.rs

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Signed-off-by: zyy17 <zyylsxm@gmail.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2025-06-23 02:56:12 +00:00
331 changed files with 14162 additions and 5470 deletions

42
.github/scripts/check-version.sh vendored Executable file
View File

@@ -0,0 +1,42 @@
#!/bin/bash
# Get current version
CURRENT_VERSION=$1
if [ -z "$CURRENT_VERSION" ]; then
echo "Error: Failed to get current version"
exit 1
fi
# Get the latest version from GitHub Releases
API_RESPONSE=$(curl -s "https://api.github.com/repos/GreptimeTeam/greptimedb/releases/latest")
if [ -z "$API_RESPONSE" ] || [ "$(echo "$API_RESPONSE" | jq -r '.message')" = "Not Found" ]; then
echo "Error: Failed to fetch latest version from GitHub"
exit 1
fi
# Get the latest version
LATEST_VERSION=$(echo "$API_RESPONSE" | jq -r '.tag_name')
if [ -z "$LATEST_VERSION" ] || [ "$LATEST_VERSION" = "null" ]; then
echo "Error: No valid version found in GitHub releases"
exit 1
fi
# Cleaned up version number format (removed possible 'v' prefix and -nightly suffix)
CLEAN_CURRENT=$(echo "$CURRENT_VERSION" | sed 's/^v//' | sed 's/-nightly-.*//')
CLEAN_LATEST=$(echo "$LATEST_VERSION" | sed 's/^v//' | sed 's/-nightly-.*//')
echo "Current version: $CLEAN_CURRENT"
echo "Latest release version: $CLEAN_LATEST"
# Use sort -V to compare versions
HIGHER_VERSION=$(printf "%s\n%s" "$CLEAN_CURRENT" "$CLEAN_LATEST" | sort -V | tail -n1)
if [ "$HIGHER_VERSION" = "$CLEAN_CURRENT" ]; then
echo "Current version ($CLEAN_CURRENT) is NEWER than or EQUAL to latest ($CLEAN_LATEST)"
echo "should-push-latest-tag=true" >> $GITHUB_OUTPUT
else
echo "Current version ($CLEAN_CURRENT) is OLDER than latest ($CLEAN_LATEST)"
echo "should-push-latest-tag=false" >> $GITHUB_OUTPUT
fi

View File

@@ -110,6 +110,8 @@ jobs:
# The 'version' use as the global tag name of the release workflow.
version: ${{ steps.create-version.outputs.version }}
should-push-latest-tag: ${{ steps.check-version.outputs.should-push-latest-tag }}
steps:
- name: Checkout
uses: actions/checkout@v4
@@ -135,6 +137,11 @@ jobs:
GITHUB_REF_NAME: ${{ github.ref_name }}
NIGHTLY_RELEASE_PREFIX: ${{ env.NIGHTLY_RELEASE_PREFIX }}
- name: Check version
id: check-version
run: |
./.github/scripts/check-version.sh "${{ steps.create-version.outputs.version }}"
- name: Allocate linux-amd64 runner
if: ${{ inputs.build_linux_amd64_artifacts || github.event_name == 'push' || github.event_name == 'schedule' }}
uses: ./.github/actions/start-runner
@@ -314,7 +321,7 @@ jobs:
image-registry-username: ${{ secrets.DOCKERHUB_USERNAME }}
image-registry-password: ${{ secrets.DOCKERHUB_TOKEN }}
version: ${{ needs.allocate-runners.outputs.version }}
push-latest-tag: ${{ github.ref_type == 'tag' && !contains(github.ref_name, 'nightly') && github.event_name != 'schedule' }}
push-latest-tag: ${{ needs.allocate-runners.outputs.should-push-latest-tag == 'true' && github.ref_type == 'tag' && !contains(github.ref_name, 'nightly') && github.event_name != 'schedule' }}
- name: Set build image result
id: set-build-image-result
@@ -332,7 +339,7 @@ jobs:
build-windows-artifacts,
release-images-to-dockerhub,
]
runs-on: ubuntu-latest
runs-on: ubuntu-latest-16-cores
# 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.
@@ -361,7 +368,7 @@ jobs:
dev-mode: false
upload-to-s3: true
update-version-info: true
push-latest-tag: ${{ github.ref_type == 'tag' && !contains(github.ref_name, 'nightly') && github.event_name != 'schedule' }}
push-latest-tag: ${{ needs.allocate-runners.outputs.should-push-latest-tag == 'true' && github.ref_type == 'tag' && !contains(github.ref_name, 'nightly') && github.event_name != 'schedule' }}
publish-github-release:
name: Create GitHub release and upload artifacts

View File

@@ -11,17 +11,17 @@ concurrency:
group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
cancel-in-progress: true
permissions:
issues: write
contents: write
pull-requests: write
jobs:
check:
runs-on: ubuntu-latest
permissions:
pull-requests: write # Add permissions to modify PRs
issues: write
timeout-minutes: 10
steps:
- uses: actions/checkout@v4
with:
persist-credentials: false
- uses: ./.github/actions/setup-cyborg
- name: Check Pull Request
working-directory: cyborg

296
Cargo.lock generated
View File

@@ -211,7 +211,7 @@ checksum = "d301b3b94cb4b2f23d7917810addbbaff90738e0ca2be692bd027e70d7e0330c"
[[package]]
name = "api"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"common-base",
"common-decimal",
@@ -944,7 +944,7 @@ dependencies = [
[[package]]
name = "auth"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"async-trait",
@@ -1586,7 +1586,7 @@ dependencies = [
[[package]]
name = "cache"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"catalog",
"common-error",
@@ -1610,7 +1610,7 @@ checksum = "37b2a672a2cb129a2e41c10b1224bb368f9f37a2b16b612598138befd7b37eb5"
[[package]]
name = "catalog"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"arrow 54.2.1",
@@ -1948,7 +1948,7 @@ checksum = "1462739cb27611015575c0c11df5df7601141071f07518d56fcc1be504cbec97"
[[package]]
name = "cli"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-stream",
"async-trait",
@@ -1975,7 +1975,6 @@ dependencies = [
"common-version",
"common-wal",
"datatypes",
"either",
"etcd-client",
"futures",
"humantime",
@@ -1986,14 +1985,14 @@ dependencies = [
"operator",
"query",
"rand 0.9.0",
"reqwest",
"reqwest 0.12.9",
"serde",
"serde_json",
"servers",
"session",
"snafu 0.8.5",
"store-api",
"substrait 0.15.0",
"substrait 0.16.0",
"table",
"tempfile",
"tokio",
@@ -2002,7 +2001,7 @@ dependencies = [
[[package]]
name = "client"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"arc-swap",
@@ -2032,7 +2031,7 @@ dependencies = [
"rand 0.9.0",
"serde_json",
"snafu 0.8.5",
"substrait 0.15.0",
"substrait 0.16.0",
"substrait 0.37.3",
"tokio",
"tokio-stream",
@@ -2073,7 +2072,7 @@ dependencies = [
[[package]]
name = "cmd"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-trait",
"auth",
@@ -2103,7 +2102,6 @@ dependencies = [
"common-wal",
"datanode",
"datatypes",
"either",
"etcd-client",
"file-engine",
"flow",
@@ -2118,13 +2116,14 @@ dependencies = [
"mito2",
"moka",
"nu-ansi-term",
"object-store",
"plugins",
"prometheus",
"prost 0.13.5",
"query",
"rand 0.9.0",
"regex",
"reqwest",
"reqwest 0.12.9",
"rexpect",
"serde",
"serde_json",
@@ -2134,7 +2133,7 @@ dependencies = [
"snafu 0.8.5",
"stat",
"store-api",
"substrait 0.15.0",
"substrait 0.16.0",
"table",
"temp-env",
"tempfile",
@@ -2181,7 +2180,7 @@ checksum = "55b672471b4e9f9e95499ea597ff64941a309b2cdbffcc46f2cc5e2d971fd335"
[[package]]
name = "common-base"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"anymap2",
"async-trait",
@@ -2203,11 +2202,11 @@ dependencies = [
[[package]]
name = "common-catalog"
version = "0.15.0"
version = "0.16.0"
[[package]]
name = "common-config"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"common-base",
"common-error",
@@ -2220,6 +2219,7 @@ dependencies = [
"humantime-serde",
"meta-client",
"num_cpus",
"object-store",
"serde",
"serde_json",
"serde_with",
@@ -2232,7 +2232,7 @@ dependencies = [
[[package]]
name = "common-datasource"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"arrow 54.2.1",
"arrow-schema 54.3.1",
@@ -2269,7 +2269,7 @@ dependencies = [
[[package]]
name = "common-decimal"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"bigdecimal 0.4.8",
"common-error",
@@ -2282,7 +2282,7 @@ dependencies = [
[[package]]
name = "common-error"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"common-macro",
"http 1.1.0",
@@ -2293,7 +2293,7 @@ dependencies = [
[[package]]
name = "common-frontend"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-trait",
"common-error",
@@ -2309,7 +2309,7 @@ dependencies = [
[[package]]
name = "common-function"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"ahash 0.8.11",
"api",
@@ -2332,6 +2332,7 @@ dependencies = [
"datafusion",
"datafusion-common",
"datafusion-expr",
"datafusion-functions-aggregate-common",
"datatypes",
"derive_more",
"geo",
@@ -2362,7 +2363,7 @@ dependencies = [
[[package]]
name = "common-greptimedb-telemetry"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-trait",
"common-runtime",
@@ -2370,7 +2371,7 @@ dependencies = [
"common-test-util",
"common-version",
"hyper 0.14.30",
"reqwest",
"reqwest 0.12.9",
"serde",
"tempfile",
"tokio",
@@ -2379,7 +2380,7 @@ dependencies = [
[[package]]
name = "common-grpc"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"arrow-flight",
@@ -2411,7 +2412,7 @@ dependencies = [
[[package]]
name = "common-grpc-expr"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"common-base",
@@ -2430,7 +2431,7 @@ dependencies = [
[[package]]
name = "common-macro"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"arc-swap",
"common-query",
@@ -2444,7 +2445,7 @@ dependencies = [
[[package]]
name = "common-mem-prof"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"anyhow",
"common-error",
@@ -2460,7 +2461,7 @@ dependencies = [
[[package]]
name = "common-meta"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"anymap2",
"api",
@@ -2525,7 +2526,7 @@ dependencies = [
[[package]]
name = "common-options"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"common-grpc",
"humantime-serde",
@@ -2534,11 +2535,11 @@ dependencies = [
[[package]]
name = "common-plugins"
version = "0.15.0"
version = "0.16.0"
[[package]]
name = "common-pprof"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"common-error",
"common-macro",
@@ -2550,7 +2551,7 @@ dependencies = [
[[package]]
name = "common-procedure"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-stream",
"async-trait",
@@ -2577,7 +2578,7 @@ dependencies = [
[[package]]
name = "common-procedure-test"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-trait",
"common-procedure",
@@ -2586,7 +2587,7 @@ dependencies = [
[[package]]
name = "common-query"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"async-trait",
@@ -2612,7 +2613,7 @@ dependencies = [
[[package]]
name = "common-recordbatch"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"arc-swap",
"common-error",
@@ -2632,7 +2633,7 @@ dependencies = [
[[package]]
name = "common-runtime"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-trait",
"clap 4.5.19",
@@ -2662,14 +2663,32 @@ dependencies = [
[[package]]
name = "common-session"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"strum 0.27.1",
]
[[package]]
name = "common-sql"
version = "0.16.0"
dependencies = [
"common-base",
"common-datasource",
"common-decimal",
"common-error",
"common-macro",
"common-time",
"datafusion-sql",
"datatypes",
"hex",
"jsonb",
"snafu 0.8.5",
"sqlparser 0.54.0 (git+https://github.com/GreptimeTeam/sqlparser-rs.git?rev=0cf6c04490d59435ee965edd2078e8855bd8471e)",
]
[[package]]
name = "common-telemetry"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"backtrace",
"common-error",
@@ -2696,7 +2715,7 @@ dependencies = [
[[package]]
name = "common-test-util"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"client",
"common-grpc",
@@ -2709,7 +2728,7 @@ dependencies = [
[[package]]
name = "common-time"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"arrow 54.2.1",
"chrono",
@@ -2727,7 +2746,7 @@ dependencies = [
[[package]]
name = "common-version"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"build-data",
"const_format",
@@ -2737,7 +2756,7 @@ dependencies = [
[[package]]
name = "common-wal"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"common-base",
"common-error",
@@ -2760,7 +2779,7 @@ dependencies = [
[[package]]
name = "common-workload"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"common-telemetry",
@@ -3716,7 +3735,7 @@ dependencies = [
[[package]]
name = "datanode"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"arrow-flight",
@@ -3762,14 +3781,14 @@ dependencies = [
"prometheus",
"prost 0.13.5",
"query",
"reqwest",
"reqwest 0.12.9",
"serde",
"serde_json",
"servers",
"session",
"snafu 0.8.5",
"store-api",
"substrait 0.15.0",
"substrait 0.16.0",
"table",
"tokio",
"toml 0.8.19",
@@ -3778,7 +3797,7 @@ dependencies = [
[[package]]
name = "datatypes"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"arrow 54.2.1",
"arrow-array 54.2.1",
@@ -4198,9 +4217,9 @@ dependencies = [
[[package]]
name = "either"
version = "1.13.0"
version = "1.15.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "60b1af1c220855b6ceac025d3f6ecdd2b7c4894bfe9cd9bda4fbb4bc7c0d4cf0"
checksum = "48c757948c5ede0e46177b7add2e67155f70e33c07fea8284df6576da70b3719"
dependencies = [
"serde",
]
@@ -4438,7 +4457,7 @@ checksum = "e8c02a5121d4ea3eb16a80748c74f5549a5665e4c21333c6098f283870fbdea6"
[[package]]
name = "file-engine"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"async-trait",
@@ -4575,7 +4594,7 @@ checksum = "8bf7cc16383c4b8d58b9905a8509f02926ce3058053c056376248d958c9df1e8"
[[package]]
name = "flow"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"arrow 54.2.1",
@@ -4640,7 +4659,7 @@ dependencies = [
"sql",
"store-api",
"strum 0.27.1",
"substrait 0.15.0",
"substrait 0.16.0",
"table",
"tokio",
"tonic 0.12.3",
@@ -4695,10 +4714,11 @@ checksum = "6c2141d6d6c8512188a7891b4b01590a45f6dac67afb4f255c4124dbb86d4eaa"
[[package]]
name = "frontend"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"arc-swap",
"async-stream",
"async-trait",
"auth",
"bytes",
@@ -4754,7 +4774,7 @@ dependencies = [
"sqlparser 0.54.0 (git+https://github.com/GreptimeTeam/sqlparser-rs.git?rev=0cf6c04490d59435ee965edd2078e8855bd8471e)",
"store-api",
"strfmt",
"substrait 0.15.0",
"substrait 0.16.0",
"table",
"tokio",
"tokio-util",
@@ -5144,7 +5164,7 @@ dependencies = [
[[package]]
name = "greptime-proto"
version = "0.1.0"
source = "git+https://github.com/GreptimeTeam/greptime-proto.git?rev=82fe5c6282f623c185b86f03e898ee8952e50cf9#82fe5c6282f623c185b86f03e898ee8952e50cf9"
source = "git+https://github.com/GreptimeTeam/greptime-proto.git?rev=ceb1af4fa9309ce65bda0367db7b384df2bb4d4f#ceb1af4fa9309ce65bda0367db7b384df2bb4d4f"
dependencies = [
"prost 0.13.5",
"serde",
@@ -5915,7 +5935,7 @@ dependencies = [
[[package]]
name = "index"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-trait",
"asynchronous-codec",
@@ -6695,7 +6715,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4979f22fdb869068da03c9f7528f8297c6fd2606bc3a4affe42e6a823fdb8da4"
dependencies = [
"cfg-if",
"windows-targets 0.48.5",
"windows-targets 0.52.6",
]
[[package]]
@@ -6800,7 +6820,7 @@ checksum = "a7a70ba024b9dc04c27ea2f0c0548feb474ec5c54bba33a7f72f873a39d07b24"
[[package]]
name = "log-query"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"chrono",
"common-error",
@@ -6812,7 +6832,7 @@ dependencies = [
[[package]]
name = "log-store"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-stream",
"async-trait",
@@ -7110,7 +7130,7 @@ dependencies = [
[[package]]
name = "meta-client"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"async-trait",
@@ -7138,7 +7158,7 @@ dependencies = [
[[package]]
name = "meta-srv"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"async-trait",
@@ -7168,6 +7188,7 @@ dependencies = [
"deadpool",
"deadpool-postgres",
"derive_builder 0.20.1",
"either",
"etcd-client",
"futures",
"h2 0.3.26",
@@ -7229,7 +7250,7 @@ dependencies = [
[[package]]
name = "metric-engine"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"aquamarine",
@@ -7239,6 +7260,7 @@ dependencies = [
"common-base",
"common-error",
"common-macro",
"common-meta",
"common-query",
"common-recordbatch",
"common-runtime",
@@ -7319,7 +7341,7 @@ dependencies = [
[[package]]
name = "mito-codec"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"bytes",
@@ -7342,7 +7364,7 @@ dependencies = [
[[package]]
name = "mito2"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"aquamarine",
@@ -8092,18 +8114,25 @@ dependencies = [
[[package]]
name = "object-store"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"anyhow",
"bytes",
"common-base",
"common-error",
"common-macro",
"common-telemetry",
"common-test-util",
"futures",
"humantime-serde",
"lazy_static",
"md5",
"moka",
"opendal",
"prometheus",
"reqwest 0.12.9",
"serde",
"snafu 0.8.5",
"tokio",
"uuid",
]
@@ -8238,7 +8267,7 @@ dependencies = [
"prometheus",
"quick-xml 0.36.2",
"reqsign",
"reqwest",
"reqwest 0.12.9",
"serde",
"serde_json",
"sha2",
@@ -8310,6 +8339,19 @@ dependencies = [
"tracing",
]
[[package]]
name = "opentelemetry-http"
version = "0.10.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7f51189ce8be654f9b5f7e70e49967ed894e84a06fc35c6c042e64ac1fc5399e"
dependencies = [
"async-trait",
"bytes",
"http 0.2.12",
"opentelemetry 0.21.0",
"reqwest 0.11.27",
]
[[package]]
name = "opentelemetry-otlp"
version = "0.14.0"
@@ -8320,10 +8362,12 @@ dependencies = [
"futures-core",
"http 0.2.12",
"opentelemetry 0.21.0",
"opentelemetry-http",
"opentelemetry-proto 0.4.0",
"opentelemetry-semantic-conventions",
"opentelemetry_sdk 0.21.2",
"prost 0.11.9",
"reqwest 0.11.27",
"thiserror 1.0.64",
"tokio",
"tonic 0.9.2",
@@ -8406,7 +8450,7 @@ dependencies = [
[[package]]
name = "operator"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"ahash 0.8.11",
"api",
@@ -8431,6 +8475,7 @@ dependencies = [
"common-query",
"common-recordbatch",
"common-runtime",
"common-sql",
"common-telemetry",
"common-test-util",
"common-time",
@@ -8461,7 +8506,7 @@ dependencies = [
"sql",
"sqlparser 0.54.0 (git+https://github.com/GreptimeTeam/sqlparser-rs.git?rev=0cf6c04490d59435ee965edd2078e8855bd8471e)",
"store-api",
"substrait 0.15.0",
"substrait 0.16.0",
"table",
"tokio",
"tokio-util",
@@ -8728,7 +8773,7 @@ dependencies = [
[[package]]
name = "partition"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"async-trait",
@@ -9016,7 +9061,7 @@ checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184"
[[package]]
name = "pipeline"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"ahash 0.8.11",
"api",
@@ -9159,7 +9204,7 @@ dependencies = [
[[package]]
name = "plugins"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"auth",
"clap 4.5.19",
@@ -9472,7 +9517,7 @@ dependencies = [
[[package]]
name = "promql"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"ahash 0.8.11",
"async-trait",
@@ -9754,7 +9799,7 @@ dependencies = [
[[package]]
name = "puffin"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-compression 0.4.13",
"async-trait",
@@ -9796,7 +9841,7 @@ dependencies = [
[[package]]
name = "query"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"ahash 0.8.11",
"api",
@@ -9862,7 +9907,7 @@ dependencies = [
"sqlparser 0.54.0 (git+https://github.com/GreptimeTeam/sqlparser-rs.git?rev=0cf6c04490d59435ee965edd2078e8855bd8471e)",
"statrs",
"store-api",
"substrait 0.15.0",
"substrait 0.16.0",
"table",
"tokio",
"tokio-stream",
@@ -10310,7 +10355,7 @@ dependencies = [
"percent-encoding",
"quick-xml 0.35.0",
"rand 0.8.5",
"reqwest",
"reqwest 0.12.9",
"rsa",
"rust-ini 0.21.1",
"serde",
@@ -10319,6 +10364,42 @@ dependencies = [
"sha2",
]
[[package]]
name = "reqwest"
version = "0.11.27"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "dd67538700a17451e7cba03ac727fb961abb7607553461627b97de0b89cf4a62"
dependencies = [
"base64 0.21.7",
"bytes",
"encoding_rs",
"futures-core",
"futures-util",
"h2 0.3.26",
"http 0.2.12",
"http-body 0.4.6",
"hyper 0.14.30",
"ipnet",
"js-sys",
"log",
"mime",
"once_cell",
"percent-encoding",
"pin-project-lite",
"serde",
"serde_json",
"serde_urlencoded",
"sync_wrapper 0.1.2",
"system-configuration",
"tokio",
"tower-service",
"url",
"wasm-bindgen",
"wasm-bindgen-futures",
"web-sys",
"winreg",
]
[[package]]
name = "reqwest"
version = "0.12.9"
@@ -11148,7 +11229,7 @@ dependencies = [
[[package]]
name = "servers"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"ahash 0.8.11",
"api",
@@ -11181,6 +11262,7 @@ dependencies = [
"common-recordbatch",
"common-runtime",
"common-session",
"common-sql",
"common-telemetry",
"common-test-util",
"common-time",
@@ -11234,7 +11316,7 @@ dependencies = [
"quoted-string",
"rand 0.9.0",
"regex",
"reqwest",
"reqwest 0.12.9",
"rust-embed",
"rustls",
"rustls-pemfile",
@@ -11269,7 +11351,7 @@ dependencies = [
[[package]]
name = "session"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"arc-swap",
@@ -11608,7 +11690,7 @@ dependencies = [
[[package]]
name = "sql"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"chrono",
@@ -11619,6 +11701,7 @@ dependencies = [
"common-error",
"common-macro",
"common-query",
"common-sql",
"common-time",
"datafusion",
"datafusion-common",
@@ -11663,7 +11746,7 @@ dependencies = [
[[package]]
name = "sqlness-runner"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-trait",
"clap 4.5.19",
@@ -11678,7 +11761,7 @@ dependencies = [
"local-ip-address",
"mysql",
"num_cpus",
"reqwest",
"reqwest 0.12.9",
"serde",
"serde_json",
"sha2",
@@ -11963,7 +12046,7 @@ dependencies = [
[[package]]
name = "stat"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"nix 0.30.1",
]
@@ -11989,7 +12072,7 @@ dependencies = [
[[package]]
name = "store-api"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"aquamarine",
@@ -12150,7 +12233,7 @@ dependencies = [
[[package]]
name = "substrait"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"async-trait",
"bytes",
@@ -12328,9 +12411,30 @@ dependencies = [
"nom",
]
[[package]]
name = "system-configuration"
version = "0.5.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ba3a3adc5c275d719af8cb4272ea1c4a6d668a777f37e115f6d11ddbc1c8e0e7"
dependencies = [
"bitflags 1.3.2",
"core-foundation",
"system-configuration-sys",
]
[[package]]
name = "system-configuration-sys"
version = "0.5.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a75fb188eb626b924683e3b95e3a48e63551fcfb51949de2f06a9d91dbee93c9"
dependencies = [
"core-foundation-sys",
"libc",
]
[[package]]
name = "table"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"async-trait",
@@ -12591,7 +12695,7 @@ checksum = "3369f5ac52d5eb6ab48c6b4ffdc8efbcad6b89c765749064ba298f2c68a16a76"
[[package]]
name = "tests-fuzz"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"arbitrary",
"async-trait",
@@ -12618,7 +12722,7 @@ dependencies = [
"paste",
"rand 0.9.0",
"rand_chacha 0.9.0",
"reqwest",
"reqwest 0.12.9",
"schemars",
"serde",
"serde_json",
@@ -12635,7 +12739,7 @@ dependencies = [
[[package]]
name = "tests-integration"
version = "0.15.0"
version = "0.16.0"
dependencies = [
"api",
"arrow-flight",
@@ -12702,7 +12806,7 @@ dependencies = [
"sql",
"sqlx",
"store-api",
"substrait 0.15.0",
"substrait 0.16.0",
"table",
"tempfile",
"time",
@@ -14501,6 +14605,16 @@ dependencies = [
"memchr",
]
[[package]]
name = "winreg"
version = "0.50.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "524e57b2c537c0f9b1e69f1965311ec12182b4122e45035b1508cd24d2adadb1"
dependencies = [
"cfg-if",
"windows-sys 0.48.0",
]
[[package]]
name = "wit-bindgen-rt"
version = "0.39.0"

View File

@@ -30,6 +30,7 @@ members = [
"src/common/recordbatch",
"src/common/runtime",
"src/common/session",
"src/common/sql",
"src/common/stat",
"src/common/substrait",
"src/common/telemetry",
@@ -71,11 +72,13 @@ members = [
resolver = "2"
[workspace.package]
version = "0.15.0"
version = "0.16.0"
edition = "2021"
license = "Apache-2.0"
[workspace.lints]
clippy.print_stdout = "warn"
clippy.print_stderr = "warn"
clippy.dbg_macro = "warn"
clippy.implicit_clone = "warn"
clippy.result_large_err = "allow"
@@ -121,6 +124,7 @@ datafusion = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "
datafusion-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-functions = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-functions-aggregate-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-optimizer = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-physical-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-physical-plan = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
@@ -130,11 +134,12 @@ deadpool = "0.12"
deadpool-postgres = "0.14"
derive_builder = "0.20"
dotenv = "0.15"
either = "1.15"
etcd-client = "0.14"
fst = "0.4.7"
futures = "0.3"
futures-util = "0.3"
greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "82fe5c6282f623c185b86f03e898ee8952e50cf9" }
greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "ceb1af4fa9309ce65bda0367db7b384df2bb4d4f" }
hex = "0.4"
http = "1"
humantime = "2.1"
@@ -258,6 +263,7 @@ common-query = { path = "src/common/query" }
common-recordbatch = { path = "src/common/recordbatch" }
common-runtime = { path = "src/common/runtime" }
common-session = { path = "src/common/session" }
common-sql = { path = "src/common/sql" }
common-telemetry = { path = "src/common/telemetry" }
common-test-util = { path = "src/common/test-util" }
common-time = { path = "src/common/time" }

View File

@@ -75,9 +75,9 @@
| --------- | ----------- |
| [Unified Observability Data](https://docs.greptime.com/user-guide/concepts/why-greptimedb) | Store metrics, logs, and traces as timestamped, contextual wide events. Query via [SQL](https://docs.greptime.com/user-guide/query-data/sql), [PromQL](https://docs.greptime.com/user-guide/query-data/promql), and [streaming](https://docs.greptime.com/user-guide/flow-computation/overview). |
| [High Performance & Cost Effective](https://docs.greptime.com/user-guide/manage-data/data-index) | Written in Rust, with a distributed query engine, [rich indexing](https://docs.greptime.com/user-guide/manage-data/data-index), and optimized columnar storage, delivering sub-second responses at PB scale. |
| [Cloud-Native Architecture](https://docs.greptime.com/user-guide/concepts/architecture) | Designed for [Kubernetes](https://docs.greptime.com/user-guide/deployments/deploy-on-kubernetes/greptimedb-operator-management), with compute/storage separation, native object storage (AWS S3, Azure Blob, etc.) and seamless cross-cloud access. |
| [Cloud-Native Architecture](https://docs.greptime.com/user-guide/concepts/architecture) | Designed for [Kubernetes](https://docs.greptime.com/user-guide/deployments-administration/deploy-on-kubernetes/greptimedb-operator-management), with compute/storage separation, native object storage (AWS S3, Azure Blob, etc.) and seamless cross-cloud access. |
| [Developer-Friendly](https://docs.greptime.com/user-guide/protocols/overview) | Access via SQL/PromQL interfaces, REST API, MySQL/PostgreSQL protocols, and popular ingestion [protocols](https://docs.greptime.com/user-guide/protocols/overview). |
| [Flexible Deployment](https://docs.greptime.com/user-guide/deployments/overview) | Deploy anywhere: edge (including ARM/[Android](https://docs.greptime.com/user-guide/deployments/run-on-android)) or cloud, with unified APIs and efficient data sync. |
| [Flexible Deployment](https://docs.greptime.com/user-guide/deployments-administration/overview) | Deploy anywhere: edge (including ARM/[Android](https://docs.greptime.com/user-guide/deployments-administration/run-on-android)) or cloud, with unified APIs and efficient data sync. |
Learn more in [Why GreptimeDB](https://docs.greptime.com/user-guide/concepts/why-greptimedb) and [Observability 2.0 and the Database for It](https://greptime.com/blogs/2025-04-25-greptimedb-observability2-new-database).

View File

@@ -185,10 +185,11 @@
| `logging.dir` | String | `./greptimedb_data/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
| `logging.level` | String | Unset | The log level. Can be `info`/`debug`/`warn`/`error`. |
| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
| `logging.otlp_endpoint` | String | `http://localhost:4318` | The OTLP tracing endpoint. |
| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
| `logging.max_log_files` | Integer | `720` | The maximum amount of log files. |
| `logging.otlp_export_protocol` | String | `http` | The OTLP tracing export protocol. Can be `grpc`/`http`. |
| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
| `slow_query` | -- | -- | The slow query log options. |
@@ -288,10 +289,11 @@
| `logging.dir` | String | `./greptimedb_data/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
| `logging.level` | String | Unset | The log level. Can be `info`/`debug`/`warn`/`error`. |
| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
| `logging.otlp_endpoint` | String | `http://localhost:4318` | The OTLP tracing endpoint. |
| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
| `logging.max_log_files` | Integer | `720` | The maximum amount of log files. |
| `logging.otlp_export_protocol` | String | `http` | The OTLP tracing export protocol. Can be `grpc`/`http`. |
| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
| `slow_query` | -- | -- | The slow query log options. |
@@ -323,6 +325,7 @@
| `selector` | String | `round_robin` | Datanode selector type.<br/>- `round_robin` (default value)<br/>- `lease_based`<br/>- `load_based`<br/>For details, please see "https://docs.greptime.com/developer-guide/metasrv/selector". |
| `use_memory_store` | Bool | `false` | Store data in memory. |
| `enable_region_failover` | Bool | `false` | Whether to enable region failover.<br/>This feature is only available on GreptimeDB running on cluster mode and<br/>- Using Remote WAL<br/>- Using shared storage (e.g., s3). |
| `region_failure_detector_initialization_delay` | String | `10m` | The delay before starting region failure detection.<br/>This delay helps prevent Metasrv from triggering unnecessary region failovers before all Datanodes are fully started.<br/>Especially useful when the cluster is not deployed with GreptimeDB Operator and maintenance mode is not enabled. |
| `allow_region_failover_on_local_wal` | Bool | `false` | Whether to allow region failover on local WAL.<br/>**This option is not recommended to be set to true, because it may lead to data loss during failover.** |
| `node_max_idle_time` | String | `24hours` | Max allowed idle time before removing node info from metasrv memory. |
| `enable_telemetry` | Bool | `true` | Whether to enable greptimedb telemetry. Enabled by default. |
@@ -370,10 +373,11 @@
| `logging.dir` | String | `./greptimedb_data/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
| `logging.level` | String | Unset | The log level. Can be `info`/`debug`/`warn`/`error`. |
| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
| `logging.otlp_endpoint` | String | `http://localhost:4318` | The OTLP tracing endpoint. |
| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
| `logging.max_log_files` | Integer | `720` | The maximum amount of log files. |
| `logging.otlp_export_protocol` | String | `http` | The OTLP tracing export protocol. Can be `grpc`/`http`. |
| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
| `export_metrics` | -- | -- | The metasrv can export its metrics and send to Prometheus compatible service (e.g. `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
@@ -432,8 +436,8 @@
| `wal.provider` | String | `raft_engine` | The provider of the WAL.<br/>- `raft_engine`: the wal is stored in the local file system by raft-engine.<br/>- `kafka`: it's remote wal that data is stored in Kafka. |
| `wal.dir` | String | Unset | The directory to store the WAL files.<br/>**It's only used when the provider is `raft_engine`**. |
| `wal.file_size` | String | `128MB` | The size of the WAL segment file.<br/>**It's only used when the provider is `raft_engine`**. |
| `wal.purge_threshold` | String | `1GB` | The threshold of the WAL size to trigger a flush.<br/>**It's only used when the provider is `raft_engine`**. |
| `wal.purge_interval` | String | `1m` | The interval to trigger a flush.<br/>**It's only used when the provider is `raft_engine`**. |
| `wal.purge_threshold` | String | `1GB` | The threshold of the WAL size to trigger a purge.<br/>**It's only used when the provider is `raft_engine`**. |
| `wal.purge_interval` | String | `1m` | The interval to trigger a purge.<br/>**It's only used when the provider is `raft_engine`**. |
| `wal.read_batch_size` | Integer | `128` | The read batch size.<br/>**It's only used when the provider is `raft_engine`**. |
| `wal.sync_write` | Bool | `false` | Whether to use sync write.<br/>**It's only used when the provider is `raft_engine`**. |
| `wal.enable_log_recycle` | Bool | `true` | Whether to reuse logically truncated log files.<br/>**It's only used when the provider is `raft_engine`**. |
@@ -534,10 +538,11 @@
| `logging.dir` | String | `./greptimedb_data/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
| `logging.level` | String | Unset | The log level. Can be `info`/`debug`/`warn`/`error`. |
| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
| `logging.otlp_endpoint` | String | `http://localhost:4318` | The OTLP tracing endpoint. |
| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
| `logging.max_log_files` | Integer | `720` | The maximum amount of log files. |
| `logging.otlp_export_protocol` | String | `http` | The OTLP tracing export protocol. Can be `grpc`/`http`. |
| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
| `export_metrics` | -- | -- | The datanode can export its metrics and send to Prometheus compatible service (e.g. `greptimedb` itself) from remote-write API.<br/>This is only used for `greptimedb` to export its own metrics internally. It's different from prometheus scrape. |
@@ -584,11 +589,14 @@
| `logging.dir` | String | `./greptimedb_data/logs` | The directory to store the log files. If set to empty, logs will not be written to files. |
| `logging.level` | String | Unset | The log level. Can be `info`/`debug`/`warn`/`error`. |
| `logging.enable_otlp_tracing` | Bool | `false` | Enable OTLP tracing. |
| `logging.otlp_endpoint` | String | `http://localhost:4317` | The OTLP tracing endpoint. |
| `logging.otlp_endpoint` | String | `http://localhost:4318` | The OTLP tracing endpoint. |
| `logging.append_stdout` | Bool | `true` | Whether to append logs to stdout. |
| `logging.log_format` | String | `text` | The log format. Can be `text`/`json`. |
| `logging.max_log_files` | Integer | `720` | The maximum amount of log files. |
| `logging.otlp_export_protocol` | String | `http` | The OTLP tracing export protocol. Can be `grpc`/`http`. |
| `logging.tracing_sample_ratio` | -- | -- | The percentage of tracing will be sampled and exported.<br/>Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.<br/>ratio > 1 are treated as 1. Fractions < 0 are treated as 0 |
| `logging.tracing_sample_ratio.default_ratio` | Float | `1.0` | -- |
| `tracing` | -- | -- | The tracing options. Only effect when compiled with `tokio-console` feature. |
| `tracing.tokio_console_addr` | String | Unset | The tokio console address. |
| `query` | -- | -- | -- |
| `query.parallelism` | Integer | `1` | Parallelism of the query engine for query sent by flownode.<br/>Default to 1, so it won't use too much cpu or memory |

View File

@@ -129,11 +129,11 @@ dir = "./greptimedb_data/wal"
## **It's only used when the provider is `raft_engine`**.
file_size = "128MB"
## The threshold of the WAL size to trigger a flush.
## The threshold of the WAL size to trigger a purge.
## **It's only used when the provider is `raft_engine`**.
purge_threshold = "1GB"
## The interval to trigger a flush.
## The interval to trigger a purge.
## **It's only used when the provider is `raft_engine`**.
purge_interval = "1m"
@@ -629,7 +629,7 @@ level = "info"
enable_otlp_tracing = false
## The OTLP tracing endpoint.
otlp_endpoint = "http://localhost:4317"
otlp_endpoint = "http://localhost:4318"
## Whether to append logs to stdout.
append_stdout = true
@@ -640,6 +640,9 @@ log_format = "text"
## The maximum amount of log files.
max_log_files = 720
## The OTLP tracing export protocol. Can be `grpc`/`http`.
otlp_export_protocol = "http"
## The percentage of tracing will be sampled and exported.
## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
## ratio > 1 are treated as 1. Fractions < 0 are treated as 0

View File

@@ -83,7 +83,7 @@ level = "info"
enable_otlp_tracing = false
## The OTLP tracing endpoint.
otlp_endpoint = "http://localhost:4317"
otlp_endpoint = "http://localhost:4318"
## Whether to append logs to stdout.
append_stdout = true
@@ -94,6 +94,9 @@ log_format = "text"
## The maximum amount of log files.
max_log_files = 720
## The OTLP tracing export protocol. Can be `grpc`/`http`.
otlp_export_protocol = "http"
## The percentage of tracing will be sampled and exported.
## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
## ratio > 1 are treated as 1. Fractions < 0 are treated as 0
@@ -105,3 +108,8 @@ default_ratio = 1.0
## The tokio console address.
## @toml2docs:none-default
#+ tokio_console_addr = "127.0.0.1"
[query]
## Parallelism of the query engine for query sent by flownode.
## Default to 1, so it won't use too much cpu or memory
parallelism = 1

View File

@@ -218,7 +218,7 @@ level = "info"
enable_otlp_tracing = false
## The OTLP tracing endpoint.
otlp_endpoint = "http://localhost:4317"
otlp_endpoint = "http://localhost:4318"
## Whether to append logs to stdout.
append_stdout = true
@@ -229,6 +229,9 @@ log_format = "text"
## The maximum amount of log files.
max_log_files = 720
## The OTLP tracing export protocol. Can be `grpc`/`http`.
otlp_export_protocol = "http"
## The percentage of tracing will be sampled and exported.
## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
## ratio > 1 are treated as 1. Fractions < 0 are treated as 0

View File

@@ -43,6 +43,11 @@ use_memory_store = false
## - Using shared storage (e.g., s3).
enable_region_failover = false
## The delay before starting region failure detection.
## This delay helps prevent Metasrv from triggering unnecessary region failovers before all Datanodes are fully started.
## Especially useful when the cluster is not deployed with GreptimeDB Operator and maintenance mode is not enabled.
region_failure_detector_initialization_delay = '10m'
## Whether to allow region failover on local WAL.
## **This option is not recommended to be set to true, because it may lead to data loss during failover.**
allow_region_failover_on_local_wal = false
@@ -220,7 +225,7 @@ level = "info"
enable_otlp_tracing = false
## The OTLP tracing endpoint.
otlp_endpoint = "http://localhost:4317"
otlp_endpoint = "http://localhost:4318"
## Whether to append logs to stdout.
append_stdout = true
@@ -231,6 +236,9 @@ log_format = "text"
## The maximum amount of log files.
max_log_files = 720
## The OTLP tracing export protocol. Can be `grpc`/`http`.
otlp_export_protocol = "http"
## The percentage of tracing will be sampled and exported.
## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
## ratio > 1 are treated as 1. Fractions < 0 are treated as 0

View File

@@ -720,7 +720,7 @@ level = "info"
enable_otlp_tracing = false
## The OTLP tracing endpoint.
otlp_endpoint = "http://localhost:4317"
otlp_endpoint = "http://localhost:4318"
## Whether to append logs to stdout.
append_stdout = true
@@ -731,6 +731,9 @@ log_format = "text"
## The maximum amount of log files.
max_log_files = 720
## The OTLP tracing export protocol. Can be `grpc`/`http`.
otlp_export_protocol = "http"
## The percentage of tracing will be sampled and exported.
## Valid range `[0, 1]`, 1 means all traces are sampled, 0 means all traces are not sampled, the default value is 1.
## ratio > 1 are treated as 1. Fractions < 0 are treated as 0

View File

@@ -55,12 +55,25 @@ async function main() {
await client.rest.issues.addLabels({
owner, repo, issue_number: number, labels: [labelDocsRequired],
})
// Get available assignees for the docs repo
const assigneesResponse = await docsClient.rest.issues.listAssignees({
owner: 'GreptimeTeam',
repo: 'docs',
})
const validAssignees = assigneesResponse.data.map(assignee => assignee.login)
core.info(`Available assignees: ${validAssignees.join(', ')}`)
// Check if the actor is a valid assignee, otherwise fallback to fengjiachun
const assignee = validAssignees.includes(actor) ? actor : 'fengjiachun'
core.info(`Assigning issue to: ${assignee}`)
await docsClient.rest.issues.create({
owner: 'GreptimeTeam',
repo: 'docs',
title: `Update docs for ${title}`,
body: `A document change request is generated from ${html_url}`,
assignee: actor,
assignee: assignee,
}).then((res) => {
core.info(`Created issue ${res.data}`)
})

View File

@@ -48,4 +48,4 @@ Please refer to [SQL query](./query.sql) for GreptimeDB and Clickhouse, and [que
## Addition
- You can tune GreptimeDB's configuration to get better performance.
- You can setup GreptimeDB to use S3 as storage, see [here](https://docs.greptime.com/user-guide/deployments/configuration#storage-options).
- You can setup GreptimeDB to use S3 as storage, see [here](https://docs.greptime.com/user-guide/deployments-administration/configuration#storage-options).

View File

@@ -83,7 +83,7 @@ If you use the [Helm Chart](https://github.com/GreptimeTeam/helm-charts) to depl
- `monitoring.enabled=true`: Deploys a standalone GreptimeDB instance dedicated to monitoring the cluster;
- `grafana.enabled=true`: Deploys Grafana and automatically imports the monitoring dashboard;
The standalone GreptimeDB instance will collect metrics from your cluster, and the dashboard will be available in the Grafana UI. For detailed deployment instructions, please refer to our [Kubernetes deployment guide](https://docs.greptime.com/user-guide/deployments-administration/deploy-on-kubernetes/getting-started).
The standalone GreptimeDB instance will collect metrics from your cluster, and the dashboard will be available in the Grafana UI. For detailed deployment instructions, please refer to our [Kubernetes deployment guide](https://docs.greptime.com/user-guide/deployments-administration-administration/deploy-on-kubernetes/getting-started).
### Self-host Prometheus and import dashboards manually

View File

@@ -31,8 +31,10 @@ excludes = [
"src/operator/src/expr_helper/trigger.rs",
"src/sql/src/statements/create/trigger.rs",
"src/sql/src/statements/show/trigger.rs",
"src/sql/src/statements/drop/trigger.rs",
"src/sql/src/parsers/create_parser/trigger.rs",
"src/sql/src/parsers/show_parser/trigger.rs",
"src/mito2/src/extension.rs",
]
[properties]

View File

@@ -22,6 +22,7 @@ use greptime_proto::v1::region::RegionResponse as RegionResponseV1;
pub struct RegionResponse {
pub affected_rows: AffectedRows,
pub extensions: HashMap<String, Vec<u8>>,
pub metadata: Vec<u8>,
}
impl RegionResponse {
@@ -29,6 +30,7 @@ impl RegionResponse {
Self {
affected_rows: region_response.affected_rows as _,
extensions: region_response.extensions,
metadata: region_response.metadata,
}
}
@@ -37,6 +39,16 @@ impl RegionResponse {
Self {
affected_rows,
extensions: Default::default(),
metadata: Vec::new(),
}
}
/// Creates one response with metadata.
pub fn from_metadata(metadata: Vec<u8>) -> Self {
Self {
affected_rows: 0,
extensions: Default::default(),
metadata,
}
}
}

View File

@@ -226,18 +226,20 @@ mod tests {
assert!(options.is_none());
let mut schema = ColumnSchema::new("test", ConcreteDataType::string_datatype(), true)
.with_fulltext_options(FulltextOptions {
enable: true,
analyzer: FulltextAnalyzer::English,
case_sensitive: false,
backend: FulltextBackend::Bloom,
})
.with_fulltext_options(FulltextOptions::new_unchecked(
true,
FulltextAnalyzer::English,
false,
FulltextBackend::Bloom,
10240,
0.01,
))
.unwrap();
schema.set_inverted_index(true);
let options = options_from_column_schema(&schema).unwrap();
assert_eq!(
options.options.get(FULLTEXT_GRPC_KEY).unwrap(),
"{\"enable\":true,\"analyzer\":\"English\",\"case-sensitive\":false,\"backend\":\"bloom\"}"
"{\"enable\":true,\"analyzer\":\"English\",\"case-sensitive\":false,\"backend\":\"bloom\",\"granularity\":10240,\"false-positive-rate-in-10000\":100}"
);
assert_eq!(
options.options.get(INVERTED_INDEX_GRPC_KEY).unwrap(),
@@ -247,16 +249,18 @@ mod tests {
#[test]
fn test_options_with_fulltext() {
let fulltext = FulltextOptions {
enable: true,
analyzer: FulltextAnalyzer::English,
case_sensitive: false,
backend: FulltextBackend::Bloom,
};
let fulltext = FulltextOptions::new_unchecked(
true,
FulltextAnalyzer::English,
false,
FulltextBackend::Bloom,
10240,
0.01,
);
let options = options_from_fulltext(&fulltext).unwrap().unwrap();
assert_eq!(
options.options.get(FULLTEXT_GRPC_KEY).unwrap(),
"{\"enable\":true,\"analyzer\":\"English\",\"case-sensitive\":false,\"backend\":\"bloom\"}"
"{\"enable\":true,\"analyzer\":\"English\",\"case-sensitive\":false,\"backend\":\"bloom\",\"granularity\":10240,\"false-positive-rate-in-10000\":100}"
);
}

View File

@@ -5,6 +5,7 @@ edition.workspace = true
license.workspace = true
[features]
enterprise = []
testing = []
[lints]

View File

@@ -14,9 +14,11 @@
pub use client::{CachedKvBackend, CachedKvBackendBuilder, MetaKvBackend};
mod builder;
mod client;
mod manager;
mod table_cache;
pub use builder::KvBackendCatalogManagerBuilder;
pub use manager::KvBackendCatalogManager;
pub use table_cache::{new_table_cache, TableCache, TableCacheRef};

View File

@@ -0,0 +1,131 @@
// 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 common_catalog::consts::DEFAULT_CATALOG_NAME;
use common_meta::cache::LayeredCacheRegistryRef;
use common_meta::key::flow::FlowMetadataManager;
use common_meta::key::TableMetadataManager;
use common_meta::kv_backend::KvBackendRef;
use common_procedure::ProcedureManagerRef;
use moka::sync::Cache;
use partition::manager::PartitionRuleManager;
#[cfg(feature = "enterprise")]
use crate::information_schema::InformationSchemaTableFactoryRef;
use crate::information_schema::{InformationExtensionRef, InformationSchemaProvider};
use crate::kvbackend::manager::{SystemCatalog, CATALOG_CACHE_MAX_CAPACITY};
use crate::kvbackend::KvBackendCatalogManager;
use crate::process_manager::ProcessManagerRef;
use crate::system_schema::pg_catalog::PGCatalogProvider;
pub struct KvBackendCatalogManagerBuilder {
information_extension: InformationExtensionRef,
backend: KvBackendRef,
cache_registry: LayeredCacheRegistryRef,
procedure_manager: Option<ProcedureManagerRef>,
process_manager: Option<ProcessManagerRef>,
#[cfg(feature = "enterprise")]
extra_information_table_factories:
std::collections::HashMap<String, InformationSchemaTableFactoryRef>,
}
impl KvBackendCatalogManagerBuilder {
pub fn new(
information_extension: InformationExtensionRef,
backend: KvBackendRef,
cache_registry: LayeredCacheRegistryRef,
) -> Self {
Self {
information_extension,
backend,
cache_registry,
procedure_manager: None,
process_manager: None,
#[cfg(feature = "enterprise")]
extra_information_table_factories: std::collections::HashMap::new(),
}
}
pub fn with_procedure_manager(mut self, procedure_manager: ProcedureManagerRef) -> Self {
self.procedure_manager = Some(procedure_manager);
self
}
pub fn with_process_manager(mut self, process_manager: ProcessManagerRef) -> Self {
self.process_manager = Some(process_manager);
self
}
/// Sets the extra information tables.
#[cfg(feature = "enterprise")]
pub fn with_extra_information_table_factories(
mut self,
factories: std::collections::HashMap<String, InformationSchemaTableFactoryRef>,
) -> Self {
self.extra_information_table_factories = factories;
self
}
pub fn build(self) -> Arc<KvBackendCatalogManager> {
let Self {
information_extension,
backend,
cache_registry,
procedure_manager,
process_manager,
#[cfg(feature = "enterprise")]
extra_information_table_factories,
} = self;
Arc::new_cyclic(|me| KvBackendCatalogManager {
information_extension,
partition_manager: Arc::new(PartitionRuleManager::new(
backend.clone(),
cache_registry
.get()
.expect("Failed to get table_route_cache"),
)),
table_metadata_manager: Arc::new(TableMetadataManager::new(backend.clone())),
system_catalog: SystemCatalog {
catalog_manager: me.clone(),
catalog_cache: Cache::new(CATALOG_CACHE_MAX_CAPACITY),
pg_catalog_cache: Cache::new(CATALOG_CACHE_MAX_CAPACITY),
information_schema_provider: {
let provider = InformationSchemaProvider::new(
DEFAULT_CATALOG_NAME.to_string(),
me.clone(),
Arc::new(FlowMetadataManager::new(backend.clone())),
process_manager.clone(),
backend.clone(),
);
#[cfg(feature = "enterprise")]
let provider = provider
.with_extra_table_factories(extra_information_table_factories.clone());
Arc::new(provider)
},
pg_catalog_provider: Arc::new(PGCatalogProvider::new(
DEFAULT_CATALOG_NAME.to_string(),
me.clone(),
)),
backend,
process_manager,
#[cfg(feature = "enterprise")]
extra_information_table_factories,
},
cache_registry,
procedure_manager,
})
}
}

View File

@@ -28,17 +28,18 @@ use common_meta::cache::{
use common_meta::key::catalog_name::CatalogNameKey;
use common_meta::key::flow::FlowMetadataManager;
use common_meta::key::schema_name::SchemaNameKey;
use common_meta::key::table_info::TableInfoValue;
use common_meta::key::table_info::{TableInfoManager, TableInfoValue};
use common_meta::key::table_name::TableNameKey;
use common_meta::key::{TableMetadataManager, TableMetadataManagerRef};
use common_meta::key::TableMetadataManagerRef;
use common_meta::kv_backend::KvBackendRef;
use common_procedure::ProcedureManagerRef;
use futures_util::stream::BoxStream;
use futures_util::{StreamExt, TryStreamExt};
use moka::sync::Cache;
use partition::manager::{PartitionRuleManager, PartitionRuleManagerRef};
use partition::manager::PartitionRuleManagerRef;
use session::context::{Channel, QueryContext};
use snafu::prelude::*;
use store_api::metric_engine_consts::METRIC_ENGINE_NAME;
use table::dist_table::DistTable;
use table::metadata::TableId;
use table::table::numbers::{NumbersTable, NUMBERS_TABLE_NAME};
@@ -51,6 +52,8 @@ use crate::error::{
CacheNotFoundSnafu, GetTableCacheSnafu, InvalidTableInfoInCatalogSnafu, ListCatalogsSnafu,
ListSchemasSnafu, ListTablesSnafu, Result, TableMetadataManagerSnafu,
};
#[cfg(feature = "enterprise")]
use crate::information_schema::InformationSchemaTableFactoryRef;
use crate::information_schema::{InformationExtensionRef, InformationSchemaProvider};
use crate::kvbackend::TableCacheRef;
use crate::process_manager::ProcessManagerRef;
@@ -66,60 +69,22 @@ use crate::CatalogManager;
#[derive(Clone)]
pub struct KvBackendCatalogManager {
/// Provides the extension methods for the `information_schema` tables
information_extension: InformationExtensionRef,
pub(super) information_extension: InformationExtensionRef,
/// Manages partition rules.
partition_manager: PartitionRuleManagerRef,
pub(super) partition_manager: PartitionRuleManagerRef,
/// Manages table metadata.
table_metadata_manager: TableMetadataManagerRef,
pub(super) table_metadata_manager: TableMetadataManagerRef,
/// A sub-CatalogManager that handles system tables
system_catalog: SystemCatalog,
pub(super) system_catalog: SystemCatalog,
/// Cache registry for all caches.
cache_registry: LayeredCacheRegistryRef,
pub(super) cache_registry: LayeredCacheRegistryRef,
/// Only available in `Standalone` mode.
procedure_manager: Option<ProcedureManagerRef>,
pub(super) procedure_manager: Option<ProcedureManagerRef>,
}
const CATALOG_CACHE_MAX_CAPACITY: u64 = 128;
pub(super) const CATALOG_CACHE_MAX_CAPACITY: u64 = 128;
impl KvBackendCatalogManager {
pub fn new(
information_extension: InformationExtensionRef,
backend: KvBackendRef,
cache_registry: LayeredCacheRegistryRef,
procedure_manager: Option<ProcedureManagerRef>,
process_manager: Option<ProcessManagerRef>,
) -> Arc<Self> {
Arc::new_cyclic(|me| Self {
information_extension,
partition_manager: Arc::new(PartitionRuleManager::new(
backend.clone(),
cache_registry
.get()
.expect("Failed to get table_route_cache"),
)),
table_metadata_manager: Arc::new(TableMetadataManager::new(backend.clone())),
system_catalog: SystemCatalog {
catalog_manager: me.clone(),
catalog_cache: Cache::new(CATALOG_CACHE_MAX_CAPACITY),
pg_catalog_cache: Cache::new(CATALOG_CACHE_MAX_CAPACITY),
information_schema_provider: Arc::new(InformationSchemaProvider::new(
DEFAULT_CATALOG_NAME.to_string(),
me.clone(),
Arc::new(FlowMetadataManager::new(backend.clone())),
process_manager.clone(),
)),
pg_catalog_provider: Arc::new(PGCatalogProvider::new(
DEFAULT_CATALOG_NAME.to_string(),
me.clone(),
)),
backend,
process_manager,
},
cache_registry,
procedure_manager,
})
}
pub fn view_info_cache(&self) -> Result<ViewInfoCacheRef> {
self.cache_registry.get().context(CacheNotFoundSnafu {
name: "view_info_cache",
@@ -142,6 +107,61 @@ impl KvBackendCatalogManager {
pub fn procedure_manager(&self) -> Option<ProcedureManagerRef> {
self.procedure_manager.clone()
}
// Override logical table's partition key indices with physical table's.
async fn override_logical_table_partition_key_indices(
table_route_cache: &TableRouteCacheRef,
table_info_manager: &TableInfoManager,
table: TableRef,
) -> Result<TableRef> {
// If the table is not a metric table, return the table directly.
if table.table_info().meta.engine != METRIC_ENGINE_NAME {
return Ok(table);
}
if let Some(table_route_value) = table_route_cache
.get(table.table_info().table_id())
.await
.context(TableMetadataManagerSnafu)?
&& let TableRoute::Logical(logical_route) = &*table_route_value
&& let Some(physical_table_info_value) = table_info_manager
.get(logical_route.physical_table_id())
.await
.context(TableMetadataManagerSnafu)?
{
let mut new_table_info = (*table.table_info()).clone();
// Remap partition key indices from physical table to logical table
new_table_info.meta.partition_key_indices = physical_table_info_value
.table_info
.meta
.partition_key_indices
.iter()
.filter_map(|&physical_index| {
// Get the column name from the physical table using the physical index
physical_table_info_value
.table_info
.meta
.schema
.column_schemas
.get(physical_index)
.and_then(|physical_column| {
// Find the corresponding index in the logical table schema
new_table_info
.meta
.schema
.column_index_by_name(physical_column.name.as_str())
})
})
.collect();
let new_table = DistTable::table(Arc::new(new_table_info));
return Ok(new_table);
}
Ok(table)
}
}
#[async_trait::async_trait]
@@ -268,10 +288,7 @@ impl CatalogManager for KvBackendCatalogManager {
let table_cache: TableCacheRef = self.cache_registry.get().context(CacheNotFoundSnafu {
name: "table_cache",
})?;
let table_route_cache: TableRouteCacheRef =
self.cache_registry.get().context(CacheNotFoundSnafu {
name: "table_route_cache",
})?;
let table = table_cache
.get_by_ref(&TableName {
catalog_name: catalog_name.to_string(),
@@ -281,55 +298,18 @@ impl CatalogManager for KvBackendCatalogManager {
.await
.context(GetTableCacheSnafu)?;
// Override logical table's partition key indices with physical table's.
if let Some(table) = &table
&& let Some(table_route_value) = table_route_cache
.get(table.table_info().table_id())
.await
.context(TableMetadataManagerSnafu)?
&& let TableRoute::Logical(logical_route) = &*table_route_value
&& let Some(physical_table_info_value) = self
.table_metadata_manager
.table_info_manager()
.get(logical_route.physical_table_id())
.await
.context(TableMetadataManagerSnafu)?
{
let mut new_table_info = (*table.table_info()).clone();
// Gather all column names from the logical table
let logical_column_names: std::collections::HashSet<_> = new_table_info
.meta
.schema
.column_schemas()
.iter()
.map(|col| &col.name)
.collect();
// Only preserve partition key indices where the corresponding columns exist in logical table
new_table_info.meta.partition_key_indices = physical_table_info_value
.table_info
.meta
.partition_key_indices
.iter()
.filter(|&&index| {
if let Some(physical_column) = physical_table_info_value
.table_info
.meta
.schema
.column_schemas
.get(index)
{
logical_column_names.contains(&physical_column.name)
} else {
false
}
})
.cloned()
.collect();
let new_table = DistTable::table(Arc::new(new_table_info));
return Ok(Some(new_table));
if let Some(table) = table {
let table_route_cache: TableRouteCacheRef =
self.cache_registry.get().context(CacheNotFoundSnafu {
name: "table_route_cache",
})?;
return Self::override_logical_table_partition_key_indices(
&table_route_cache,
self.table_metadata_manager.table_info_manager(),
table,
)
.await
.map(Some);
}
if channel == Channel::Postgres {
@@ -342,7 +322,7 @@ impl CatalogManager for KvBackendCatalogManager {
}
}
Ok(table)
Ok(None)
}
async fn tables_by_ids(
@@ -394,8 +374,20 @@ impl CatalogManager for KvBackendCatalogManager {
let catalog = catalog.to_string();
let schema = schema.to_string();
let semaphore = Arc::new(Semaphore::new(CONCURRENCY));
let table_route_cache: Result<TableRouteCacheRef> =
self.cache_registry.get().context(CacheNotFoundSnafu {
name: "table_route_cache",
});
common_runtime::spawn_global(async move {
let table_route_cache = match table_route_cache {
Ok(table_route_cache) => table_route_cache,
Err(e) => {
let _ = tx.send(Err(e)).await;
return;
}
};
let table_id_stream = metadata_manager
.table_name_manager()
.tables(&catalog, &schema)
@@ -422,6 +414,7 @@ impl CatalogManager for KvBackendCatalogManager {
let metadata_manager = metadata_manager.clone();
let tx = tx.clone();
let semaphore = semaphore.clone();
let table_route_cache = table_route_cache.clone();
common_runtime::spawn_global(async move {
// we don't explicitly close the semaphore so just ignore the potential error.
let _ = semaphore.acquire().await;
@@ -439,6 +432,16 @@ impl CatalogManager for KvBackendCatalogManager {
};
for table in table_info_values.into_values().map(build_table) {
let table = if let Ok(table) = table {
Self::override_logical_table_partition_key_indices(
&table_route_cache,
metadata_manager.table_info_manager(),
table,
)
.await
} else {
table
};
if tx.send(table).await.is_err() {
return;
}
@@ -468,16 +471,19 @@ fn build_table(table_info_value: TableInfoValue) -> Result<TableRef> {
/// - information_schema.{tables}
/// - pg_catalog.{tables}
#[derive(Clone)]
struct SystemCatalog {
catalog_manager: Weak<KvBackendCatalogManager>,
catalog_cache: Cache<String, Arc<InformationSchemaProvider>>,
pg_catalog_cache: Cache<String, Arc<PGCatalogProvider>>,
pub(super) struct SystemCatalog {
pub(super) catalog_manager: Weak<KvBackendCatalogManager>,
pub(super) catalog_cache: Cache<String, Arc<InformationSchemaProvider>>,
pub(super) pg_catalog_cache: Cache<String, Arc<PGCatalogProvider>>,
// system_schema_provider for default catalog
information_schema_provider: Arc<InformationSchemaProvider>,
pg_catalog_provider: Arc<PGCatalogProvider>,
backend: KvBackendRef,
process_manager: Option<ProcessManagerRef>,
pub(super) information_schema_provider: Arc<InformationSchemaProvider>,
pub(super) pg_catalog_provider: Arc<PGCatalogProvider>,
pub(super) backend: KvBackendRef,
pub(super) process_manager: Option<ProcessManagerRef>,
#[cfg(feature = "enterprise")]
pub(super) extra_information_table_factories:
std::collections::HashMap<String, InformationSchemaTableFactoryRef>,
}
impl SystemCatalog {
@@ -541,12 +547,17 @@ impl SystemCatalog {
if schema == INFORMATION_SCHEMA_NAME {
let information_schema_provider =
self.catalog_cache.get_with_by_ref(catalog, move || {
Arc::new(InformationSchemaProvider::new(
let provider = InformationSchemaProvider::new(
catalog.to_string(),
self.catalog_manager.clone(),
Arc::new(FlowMetadataManager::new(self.backend.clone())),
self.process_manager.clone(),
))
self.backend.clone(),
);
#[cfg(feature = "enterprise")]
let provider = provider
.with_extra_table_factories(self.extra_information_table_factories.clone());
Arc::new(provider)
});
information_schema_provider.table(table_name)
} else if schema == PG_CATALOG_NAME && channel == Channel::Postgres {

View File

@@ -352,11 +352,13 @@ impl MemoryCatalogManager {
}
fn create_catalog_entry(self: &Arc<Self>, catalog: String) -> SchemaEntries {
let backend = Arc::new(MemoryKvBackend::new());
let information_schema_provider = InformationSchemaProvider::new(
catalog,
Arc::downgrade(self) as Weak<dyn CatalogManager>,
Arc::new(FlowMetadataManager::new(Arc::new(MemoryKvBackend::new()))),
Arc::new(FlowMetadataManager::new(backend.clone())),
None, // we don't need ProcessManager on regions server.
backend,
);
let information_schema = information_schema_provider.tables().clone();

View File

@@ -21,7 +21,7 @@ use std::sync::{Arc, RwLock};
use api::v1::frontend::{KillProcessRequest, ListProcessRequest, ProcessInfo};
use common_base::cancellation::CancellationHandle;
use common_frontend::selector::{FrontendSelector, MetaClientSelector};
use common_telemetry::{debug, info};
use common_telemetry::{debug, info, warn};
use common_time::util::current_time_millis;
use meta_client::MetaClientRef;
use snafu::{ensure, OptionExt, ResultExt};
@@ -141,14 +141,20 @@ impl ProcessManager {
.await
.context(error::InvokeFrontendSnafu)?;
for mut f in frontends {
processes.extend(
f.list_process(ListProcessRequest {
let result = f
.list_process(ListProcessRequest {
catalog: catalog.unwrap_or_default().to_string(),
})
.await
.context(error::InvokeFrontendSnafu)?
.processes,
);
.context(error::InvokeFrontendSnafu);
match result {
Ok(resp) => {
processes.extend(resp.processes);
}
Err(e) => {
warn!(e; "Skipping failing node: {:?}", f)
}
}
}
}
processes.extend(self.local_processes(catalog)?);

View File

@@ -15,7 +15,7 @@
pub mod information_schema;
mod memory_table;
pub mod pg_catalog;
mod predicate;
pub mod predicate;
mod utils;
use std::collections::HashMap;
@@ -96,7 +96,7 @@ trait SystemSchemaProviderInner {
}
}
pub(crate) trait SystemTable {
pub trait SystemTable {
fn table_id(&self) -> TableId;
fn table_name(&self) -> &'static str;
@@ -110,7 +110,7 @@ pub(crate) trait SystemTable {
}
}
pub(crate) type SystemTableRef = Arc<dyn SystemTable + Send + Sync>;
pub type SystemTableRef = Arc<dyn SystemTable + Send + Sync>;
struct SystemTableDataSource {
table: SystemTableRef,

View File

@@ -19,7 +19,7 @@ mod information_memory_table;
pub mod key_column_usage;
mod partitions;
mod procedure_info;
mod process_list;
pub mod process_list;
pub mod region_peers;
mod region_statistics;
mod runtime_metrics;
@@ -38,6 +38,7 @@ use common_meta::cluster::NodeInfo;
use common_meta::datanode::RegionStat;
use common_meta::key::flow::flow_state::FlowStat;
use common_meta::key::flow::FlowMetadataManager;
use common_meta::kv_backend::KvBackendRef;
use common_procedure::ProcedureInfo;
use common_recordbatch::SendableRecordBatchStream;
use datatypes::schema::SchemaRef;
@@ -112,6 +113,25 @@ macro_rules! setup_memory_table {
};
}
#[cfg(feature = "enterprise")]
pub struct MakeInformationTableRequest {
pub catalog_name: String,
pub catalog_manager: Weak<dyn CatalogManager>,
pub kv_backend: KvBackendRef,
}
/// A factory trait for making information schema tables.
///
/// This trait allows for extensibility of the information schema by providing
/// a way to dynamically create custom information schema tables.
#[cfg(feature = "enterprise")]
pub trait InformationSchemaTableFactory {
fn make_information_table(&self, req: MakeInformationTableRequest) -> SystemTableRef;
}
#[cfg(feature = "enterprise")]
pub type InformationSchemaTableFactoryRef = Arc<dyn InformationSchemaTableFactory + Send + Sync>;
/// The `information_schema` tables info provider.
pub struct InformationSchemaProvider {
catalog_name: String,
@@ -119,6 +139,10 @@ pub struct InformationSchemaProvider {
process_manager: Option<ProcessManagerRef>,
flow_metadata_manager: Arc<FlowMetadataManager>,
tables: HashMap<String, TableRef>,
#[allow(dead_code)]
kv_backend: KvBackendRef,
#[cfg(feature = "enterprise")]
extra_table_factories: HashMap<String, InformationSchemaTableFactoryRef>,
}
impl SystemSchemaProvider for InformationSchemaProvider {
@@ -128,6 +152,7 @@ impl SystemSchemaProvider for InformationSchemaProvider {
&self.tables
}
}
impl SystemSchemaProviderInner for InformationSchemaProvider {
fn catalog_name(&self) -> &str {
&self.catalog_name
@@ -215,7 +240,22 @@ impl SystemSchemaProviderInner for InformationSchemaProvider {
.process_manager
.as_ref()
.map(|p| Arc::new(InformationSchemaProcessList::new(p.clone())) as _),
_ => None,
table_name => {
#[cfg(feature = "enterprise")]
return self.extra_table_factories.get(table_name).map(|factory| {
let req = MakeInformationTableRequest {
catalog_name: self.catalog_name.clone(),
catalog_manager: self.catalog_manager.clone(),
kv_backend: self.kv_backend.clone(),
};
factory.make_information_table(req)
});
#[cfg(not(feature = "enterprise"))]
{
let _ = table_name;
None
}
}
}
}
}
@@ -226,6 +266,7 @@ impl InformationSchemaProvider {
catalog_manager: Weak<dyn CatalogManager>,
flow_metadata_manager: Arc<FlowMetadataManager>,
process_manager: Option<ProcessManagerRef>,
kv_backend: KvBackendRef,
) -> Self {
let mut provider = Self {
catalog_name,
@@ -233,6 +274,9 @@ impl InformationSchemaProvider {
flow_metadata_manager,
process_manager,
tables: HashMap::new(),
kv_backend,
#[cfg(feature = "enterprise")]
extra_table_factories: HashMap::new(),
};
provider.build_tables();
@@ -240,6 +284,16 @@ impl InformationSchemaProvider {
provider
}
#[cfg(feature = "enterprise")]
pub(crate) fn with_extra_table_factories(
mut self,
factories: HashMap<String, InformationSchemaTableFactoryRef>,
) -> Self {
self.extra_table_factories = factories;
self.build_tables();
self
}
fn build_tables(&mut self) {
let mut tables = HashMap::new();
@@ -290,16 +344,19 @@ impl InformationSchemaProvider {
if let Some(process_list) = self.build_table(PROCESS_LIST) {
tables.insert(PROCESS_LIST.to_string(), process_list);
}
#[cfg(feature = "enterprise")]
for name in self.extra_table_factories.keys() {
tables.insert(name.to_string(), self.build_table(name).expect(name));
}
// Add memory tables
for name in MEMORY_TABLES.iter() {
tables.insert((*name).to_string(), self.build_table(name).expect(name));
}
self.tables = tables;
}
}
trait InformationTable {
pub trait InformationTable {
fn table_id(&self) -> TableId;
fn table_name(&self) -> &'static str;

View File

@@ -39,14 +39,14 @@ use crate::process_manager::ProcessManagerRef;
use crate::system_schema::information_schema::InformationTable;
/// Column names of `information_schema.process_list`
const ID: &str = "id";
const CATALOG: &str = "catalog";
const SCHEMAS: &str = "schemas";
const QUERY: &str = "query";
const CLIENT: &str = "client";
const FRONTEND: &str = "frontend";
const START_TIMESTAMP: &str = "start_timestamp";
const ELAPSED_TIME: &str = "elapsed_time";
pub const ID: &str = "id";
pub const CATALOG: &str = "catalog";
pub const SCHEMAS: &str = "schemas";
pub const QUERY: &str = "query";
pub const CLIENT: &str = "client";
pub const FRONTEND: &str = "frontend";
pub const START_TIMESTAMP: &str = "start_timestamp";
pub const ELAPSED_TIME: &str = "elapsed_time";
/// `information_schema.process_list` table implementation that tracks running
/// queries in current cluster.

View File

@@ -48,3 +48,4 @@ pub const FLOWS: &str = "flows";
pub const PROCEDURE_INFO: &str = "procedure_info";
pub const REGION_STATISTICS: &str = "region_statistics";
pub const PROCESS_LIST: &str = "process_list";
pub const TRIGGER_LIST: &str = "trigger_list";

View File

@@ -207,6 +207,7 @@ mod tests {
use session::context::QueryContext;
use super::*;
use crate::kvbackend::KvBackendCatalogManagerBuilder;
use crate::memory::MemoryCatalogManager;
#[test]
@@ -323,13 +324,13 @@ mod tests {
.build(),
);
let catalog_manager = KvBackendCatalogManager::new(
let catalog_manager = KvBackendCatalogManagerBuilder::new(
Arc::new(NoopInformationExtension),
backend.clone(),
layered_cache_registry,
None,
None,
);
)
.build();
let table_metadata_manager = TableMetadataManager::new(backend);
let mut view_info = common_meta::key::test_utils::new_test_table_info(1024, vec![]);
view_info.table_type = TableType::View;

View File

@@ -43,7 +43,6 @@ common-time.workspace = true
common-version.workspace = true
common-wal.workspace = true
datatypes.workspace = true
either = "1.8"
etcd-client.workspace = true
futures.workspace = true
humantime.workspace = true

View File

@@ -160,6 +160,7 @@ fn create_table_info(table_id: TableId, table_name: TableName) -> RawTableInfo {
options: Default::default(),
region_numbers: (1..=100).collect(),
partition_key_indices: vec![],
column_ids: vec![],
};
RawTableInfo {

View File

@@ -12,6 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
#![allow(clippy::print_stdout)]
mod bench;
mod data;
mod database;

View File

@@ -301,7 +301,6 @@ struct MetaInfoTool {
#[async_trait]
impl Tool for MetaInfoTool {
#[allow(clippy::print_stdout)]
async fn do_work(&self) -> std::result::Result<(), BoxedError> {
let result = MetadataSnapshotManager::info(
&self.inner,

View File

@@ -31,7 +31,7 @@ use base64::prelude::BASE64_STANDARD;
use base64::Engine;
use common_catalog::build_db_string;
use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
use common_error::ext::{BoxedError, ErrorExt};
use common_error::ext::BoxedError;
use common_grpc::flight::do_put::DoPutResponse;
use common_grpc::flight::{FlightDecoder, FlightMessage};
use common_query::Output;
@@ -48,7 +48,7 @@ use tonic::transport::Channel;
use crate::error::{
ConvertFlightDataSnafu, Error, FlightGetSnafu, IllegalFlightMessagesSnafu,
InvalidTonicMetadataValueSnafu, ServerSnafu,
InvalidTonicMetadataValueSnafu,
};
use crate::{error, from_grpc_response, Client, Result};
@@ -196,12 +196,22 @@ impl Database {
/// Retry if connection fails, max_retries is the max number of retries, so the total wait time
/// is `max_retries * GRPC_CONN_TIMEOUT`
pub async fn handle_with_retry(&self, request: Request, max_retries: u32) -> Result<u32> {
pub async fn handle_with_retry(
&self,
request: Request,
max_retries: u32,
hints: &[(&str, &str)],
) -> Result<u32> {
let mut client = make_database_client(&self.client)?.inner;
let mut retries = 0;
let request = self.to_rpc_request(request);
loop {
let raw_response = client.handle(request.clone()).await;
let mut tonic_request = tonic::Request::new(request.clone());
let metadata = tonic_request.metadata_mut();
Self::put_hints(metadata, hints)?;
let raw_response = client.handle(tonic_request).await;
match (raw_response, retries < max_retries) {
(Ok(resp), _) => return from_grpc_response(resp.into_inner()),
(Err(err), true) => {
@@ -292,21 +302,16 @@ impl Database {
let response = client.mut_inner().do_get(request).await.or_else(|e| {
let tonic_code = e.code();
let e: Error = e.into();
let code = e.status_code();
let msg = e.to_string();
let error =
Err(BoxedError::new(ServerSnafu { code, msg }.build())).with_context(|_| {
FlightGetSnafu {
addr: client.addr().to_string(),
tonic_code,
}
});
error!(
"Failed to do Flight get, addr: {}, code: {}, source: {:?}",
client.addr(),
tonic_code,
error
e
);
let error = Err(BoxedError::new(e)).with_context(|_| FlightGetSnafu {
addr: client.addr().to_string(),
tonic_code,
});
error
})?;
@@ -436,8 +441,11 @@ mod tests {
use api::v1::auth_header::AuthScheme;
use api::v1::{AuthHeader, Basic};
use common_error::status_code::StatusCode;
use tonic::{Code, Status};
use super::*;
use crate::error::TonicSnafu;
#[test]
fn test_flight_ctx() {
@@ -460,4 +468,19 @@ mod tests {
})
)
}
#[test]
fn test_from_tonic_status() {
let expected = TonicSnafu {
code: StatusCode::Internal,
msg: "blabla".to_string(),
tonic_code: Code::Internal,
}
.build();
let status = Status::new(Code::Internal, "blabla");
let actual: Error = status.into();
assert_eq!(expected.to_string(), actual.to_string());
}
}

View File

@@ -14,13 +14,13 @@
use std::any::Any;
use common_error::define_from_tonic_status;
use common_error::ext::{BoxedError, ErrorExt};
use common_error::status_code::{convert_tonic_code_to_status_code, StatusCode};
use common_error::{GREPTIME_DB_HEADER_ERROR_CODE, GREPTIME_DB_HEADER_ERROR_MSG};
use common_error::status_code::StatusCode;
use common_macro::stack_trace_debug;
use snafu::{location, Location, Snafu};
use tonic::metadata::errors::InvalidMetadataValue;
use tonic::{Code, Status};
use tonic::Code;
#[derive(Snafu)]
#[snafu(visibility(pub))]
@@ -124,6 +124,15 @@ pub enum Error {
location: Location,
source: datatypes::error::Error,
},
#[snafu(display("{}", msg))]
Tonic {
code: StatusCode,
msg: String,
tonic_code: Code,
#[snafu(implicit)]
location: Location,
},
}
pub type Result<T> = std::result::Result<T, Error>;
@@ -135,7 +144,7 @@ impl ErrorExt for Error {
| Error::MissingField { .. }
| Error::IllegalDatabaseResponse { .. } => StatusCode::Internal,
Error::Server { code, .. } => *code,
Error::Server { code, .. } | Error::Tonic { code, .. } => *code,
Error::FlightGet { source, .. }
| Error::RegionServer { source, .. }
| Error::FlowServer { source, .. } => source.status_code(),
@@ -153,34 +162,7 @@ impl ErrorExt for Error {
}
}
impl From<Status> for Error {
fn from(e: Status) -> Self {
fn get_metadata_value(e: &Status, key: &str) -> Option<String> {
e.metadata()
.get(key)
.and_then(|v| String::from_utf8(v.as_bytes().to_vec()).ok())
}
let code = get_metadata_value(&e, GREPTIME_DB_HEADER_ERROR_CODE).and_then(|s| {
if let Ok(code) = s.parse::<u32>() {
StatusCode::from_u32(code)
} else {
None
}
});
let tonic_code = e.code();
let code = code.unwrap_or_else(|| convert_tonic_code_to_status_code(tonic_code));
let msg = get_metadata_value(&e, GREPTIME_DB_HEADER_ERROR_MSG)
.unwrap_or_else(|| e.message().to_string());
Self::Server {
code,
msg,
location: location!(),
}
}
}
define_from_tonic_status!(Error, Tonic);
impl Error {
pub fn should_retry(&self) -> bool {

View File

@@ -21,7 +21,7 @@ use arc_swap::ArcSwapOption;
use arrow_flight::Ticket;
use async_stream::stream;
use async_trait::async_trait;
use common_error::ext::{BoxedError, ErrorExt};
use common_error::ext::BoxedError;
use common_error::status_code::StatusCode;
use common_grpc::flight::{FlightDecoder, FlightMessage};
use common_meta::error::{self as meta_error, Result as MetaResult};
@@ -107,24 +107,18 @@ impl RegionRequester {
.mut_inner()
.do_get(ticket)
.await
.map_err(|e| {
.or_else(|e| {
let tonic_code = e.code();
let e: error::Error = e.into();
let code = e.status_code();
let msg = e.to_string();
let error = ServerSnafu { code, msg }
.fail::<()>()
.map_err(BoxedError::new)
.with_context(|_| FlightGetSnafu {
tonic_code,
addr: flight_client.addr().to_string(),
})
.unwrap_err();
error!(
e; "Failed to do Flight get, addr: {}, code: {}",
flight_client.addr(),
tonic_code
);
let error = Err(BoxedError::new(e)).with_context(|_| FlightGetSnafu {
addr: flight_client.addr().to_string(),
tonic_code,
});
error
})?;

View File

@@ -16,7 +16,7 @@ default = [
"meta-srv/pg_kvbackend",
"meta-srv/mysql_kvbackend",
]
enterprise = ["common-meta/enterprise", "frontend/enterprise", "meta-srv/enterprise"]
enterprise = ["common-meta/enterprise", "frontend/enterprise", "meta-srv/enterprise", "catalog/enterprise"]
tokio-console = ["common-telemetry/tokio-console"]
[lints]
@@ -52,7 +52,6 @@ common-version.workspace = true
common-wal.workspace = true
datanode.workspace = true
datatypes.workspace = true
either = "1.8"
etcd-client.workspace = true
file-engine.workspace = true
flow.workspace = true
@@ -67,6 +66,7 @@ metric-engine.workspace = true
mito2.workspace = true
moka.workspace = true
nu-ansi-term = "0.46"
object-store.workspace = true
plugins.workspace = true
prometheus.workspace = true
prost.workspace = true

View File

@@ -280,7 +280,7 @@ mod tests {
use common_config::ENV_VAR_SEP;
use common_test_util::temp_dir::create_named_temp_file;
use datanode::config::{FileConfig, GcsConfig, ObjectStoreConfig, S3Config};
use object_store::config::{FileConfig, GcsConfig, ObjectStoreConfig, S3Config};
use servers::heartbeat_options::HeartbeatOptions;
use super::*;

View File

@@ -18,7 +18,7 @@ use std::time::Duration;
use cache::{build_fundamental_cache_registry, with_default_composite_cache_registry};
use catalog::information_extension::DistributedInformationExtension;
use catalog::kvbackend::{CachedKvBackendBuilder, KvBackendCatalogManager, MetaKvBackend};
use catalog::kvbackend::{CachedKvBackendBuilder, KvBackendCatalogManagerBuilder, MetaKvBackend};
use clap::Parser;
use client::client_manager::NodeClients;
use common_base::Plugins;
@@ -342,13 +342,12 @@ impl StartCommand {
let information_extension =
Arc::new(DistributedInformationExtension::new(meta_client.clone()));
let catalog_manager = KvBackendCatalogManager::new(
let catalog_manager = KvBackendCatalogManagerBuilder::new(
information_extension,
cached_meta_backend.clone(),
layered_cache_registry.clone(),
None,
None,
);
)
.build();
let table_metadata_manager =
Arc::new(TableMetadataManager::new(cached_meta_backend.clone()));
@@ -371,8 +370,11 @@ impl StartCommand {
let flow_metadata_manager = Arc::new(FlowMetadataManager::new(cached_meta_backend.clone()));
let flow_auth_header = get_flow_auth_options(&opts).context(StartFlownodeSnafu)?;
let frontend_client =
FrontendClient::from_meta_client(meta_client.clone(), flow_auth_header);
let frontend_client = FrontendClient::from_meta_client(
meta_client.clone(),
flow_auth_header,
opts.query.clone(),
);
let frontend_client = Arc::new(frontend_client);
let flownode_builder = FlownodeBuilder::new(
opts.clone(),

View File

@@ -19,7 +19,7 @@ use std::time::Duration;
use async_trait::async_trait;
use cache::{build_fundamental_cache_registry, with_default_composite_cache_registry};
use catalog::information_extension::DistributedInformationExtension;
use catalog::kvbackend::{CachedKvBackendBuilder, KvBackendCatalogManager, MetaKvBackend};
use catalog::kvbackend::{CachedKvBackendBuilder, KvBackendCatalogManagerBuilder, MetaKvBackend};
use catalog::process_manager::ProcessManager;
use clap::Parser;
use client::client_manager::NodeClients;
@@ -102,7 +102,7 @@ impl App for Instance {
#[derive(Parser)]
pub struct Command {
#[clap(subcommand)]
subcmd: SubCommand,
pub subcmd: SubCommand,
}
impl Command {
@@ -116,7 +116,7 @@ impl Command {
}
#[derive(Parser)]
enum SubCommand {
pub enum SubCommand {
Start(StartCommand),
}
@@ -153,7 +153,7 @@ pub struct StartCommand {
#[clap(long)]
postgres_addr: Option<String>,
#[clap(short, long)]
config_file: Option<String>,
pub config_file: Option<String>,
#[clap(short, long)]
influxdb_enable: Option<bool>,
#[clap(long, value_delimiter = ',', num_args = 1..)]
@@ -169,7 +169,7 @@ pub struct StartCommand {
#[clap(long)]
disable_dashboard: Option<bool>,
#[clap(long, default_value = "GREPTIMEDB_FRONTEND")]
env_prefix: String,
pub env_prefix: String,
}
impl StartCommand {
@@ -350,13 +350,20 @@ impl StartCommand {
addrs::resolve_addr(&opts.grpc.bind_addr, Some(&opts.grpc.server_addr)),
Some(meta_client.clone()),
));
let catalog_manager = KvBackendCatalogManager::new(
let builder = KvBackendCatalogManagerBuilder::new(
information_extension,
cached_meta_backend.clone(),
layered_cache_registry.clone(),
None,
Some(process_manager.clone()),
);
)
.with_process_manager(process_manager.clone());
#[cfg(feature = "enterprise")]
let builder = if let Some(factories) = plugins.get() {
builder.with_extra_information_table_factories(factories)
} else {
builder
};
let catalog_manager = builder.build();
let executor = HandlerGroupExecutor::new(vec![
Arc::new(ParseMailboxMessageHandler),

View File

@@ -54,6 +54,10 @@ impl Instance {
pub fn get_inner(&self) -> &MetasrvInstance {
&self.instance
}
pub fn mut_inner(&mut self) -> &mut MetasrvInstance {
&mut self.instance
}
}
#[async_trait]

View File

@@ -20,7 +20,7 @@ use std::{fs, path};
use async_trait::async_trait;
use cache::{build_fundamental_cache_registry, with_default_composite_cache_registry};
use catalog::information_schema::InformationExtension;
use catalog::kvbackend::KvBackendCatalogManager;
use catalog::kvbackend::KvBackendCatalogManagerBuilder;
use catalog::process_manager::ProcessManager;
use clap::Parser;
use client::api::v1::meta::RegionRole;
@@ -30,20 +30,16 @@ use common_catalog::consts::{MIN_USER_FLOW_ID, MIN_USER_TABLE_ID};
use common_config::{metadata_store_dir, Configurable, KvBackendConfig};
use common_error::ext::BoxedError;
use common_meta::cache::LayeredCacheRegistryBuilder;
use common_meta::cache_invalidator::CacheInvalidatorRef;
use common_meta::cluster::{NodeInfo, NodeStatus};
use common_meta::datanode::RegionStat;
use common_meta::ddl::flow_meta::{FlowMetadataAllocator, FlowMetadataAllocatorRef};
use common_meta::ddl::table_meta::{TableMetadataAllocator, TableMetadataAllocatorRef};
use common_meta::ddl::flow_meta::FlowMetadataAllocator;
use common_meta::ddl::table_meta::TableMetadataAllocator;
use common_meta::ddl::{DdlContext, NoopRegionFailureDetectorControl, ProcedureExecutorRef};
use common_meta::ddl_manager::DdlManager;
#[cfg(feature = "enterprise")]
use common_meta::ddl_manager::TriggerDdlManagerRef;
use common_meta::key::flow::flow_state::FlowStat;
use common_meta::key::flow::{FlowMetadataManager, FlowMetadataManagerRef};
use common_meta::key::flow::FlowMetadataManager;
use common_meta::key::{TableMetadataManager, TableMetadataManagerRef};
use common_meta::kv_backend::KvBackendRef;
use common_meta::node_manager::NodeManagerRef;
use common_meta::peer::Peer;
use common_meta::region_keeper::MemoryRegionKeeper;
use common_meta::region_registry::LeaderRegionRegistry;
@@ -261,15 +257,34 @@ pub struct Instance {
flownode: FlownodeInstance,
procedure_manager: ProcedureManagerRef,
wal_options_allocator: WalOptionsAllocatorRef,
// The components of standalone, which make it easier to expand based
// on the components.
#[cfg(feature = "enterprise")]
components: Components,
// Keep the logging guard to prevent the worker from being dropped.
_guard: Vec<WorkerGuard>,
}
#[cfg(feature = "enterprise")]
pub struct Components {
pub plugins: Plugins,
pub kv_backend: KvBackendRef,
pub frontend_client: Arc<FrontendClient>,
pub catalog_manager: catalog::CatalogManagerRef,
}
impl Instance {
/// Find the socket addr of a server by its `name`.
pub fn server_addr(&self, name: &str) -> Option<SocketAddr> {
self.frontend.server_handlers().addr(name)
}
#[cfg(feature = "enterprise")]
pub fn components(&self) -> &Components {
&self.components
}
}
#[async_trait]
@@ -529,13 +544,20 @@ impl StartCommand {
));
let process_manager = Arc::new(ProcessManager::new(opts.grpc.server_addr.clone(), None));
let catalog_manager = KvBackendCatalogManager::new(
let builder = KvBackendCatalogManagerBuilder::new(
information_extension.clone(),
kv_backend.clone(),
layered_cache_registry.clone(),
Some(procedure_manager.clone()),
Some(process_manager.clone()),
);
)
.with_procedure_manager(procedure_manager.clone())
.with_process_manager(process_manager.clone());
#[cfg(feature = "enterprise")]
let builder = if let Some(factories) = plugins.get() {
builder.with_extra_information_table_factories(factories)
} else {
builder
};
let catalog_manager = builder.build();
let table_metadata_manager =
Self::create_table_metadata_manager(kv_backend.clone()).await?;
@@ -549,14 +571,15 @@ impl StartCommand {
// for standalone not use grpc, but get a handler to frontend grpc client without
// actually make a connection
let (frontend_client, frontend_instance_handler) =
FrontendClient::from_empty_grpc_handler();
FrontendClient::from_empty_grpc_handler(opts.query.clone());
let frontend_client = Arc::new(frontend_client);
let flow_builder = FlownodeBuilder::new(
flownode_options,
plugins.clone(),
table_metadata_manager.clone(),
catalog_manager.clone(),
flow_metadata_manager.clone(),
Arc::new(frontend_client.clone()),
frontend_client.clone(),
);
let flownode = flow_builder
.build()
@@ -594,28 +617,36 @@ impl StartCommand {
.await
.context(error::BuildWalOptionsAllocatorSnafu)?;
let wal_options_allocator = Arc::new(wal_options_allocator);
let table_meta_allocator = Arc::new(TableMetadataAllocator::new(
let table_metadata_allocator = Arc::new(TableMetadataAllocator::new(
table_id_sequence,
wal_options_allocator.clone(),
));
let flow_meta_allocator = Arc::new(FlowMetadataAllocator::with_noop_peer_allocator(
let flow_metadata_allocator = Arc::new(FlowMetadataAllocator::with_noop_peer_allocator(
flow_id_sequence,
));
let ddl_context = DdlContext {
node_manager: node_manager.clone(),
cache_invalidator: layered_cache_registry.clone(),
memory_region_keeper: Arc::new(MemoryRegionKeeper::default()),
leader_region_registry: Arc::new(LeaderRegionRegistry::default()),
table_metadata_manager: table_metadata_manager.clone(),
table_metadata_allocator: table_metadata_allocator.clone(),
flow_metadata_manager: flow_metadata_manager.clone(),
flow_metadata_allocator: flow_metadata_allocator.clone(),
region_failure_detector_controller: Arc::new(NoopRegionFailureDetectorControl),
};
let procedure_manager_c = procedure_manager.clone();
let ddl_manager = DdlManager::try_new(ddl_context, procedure_manager_c, true)
.context(error::InitDdlManagerSnafu)?;
#[cfg(feature = "enterprise")]
let trigger_ddl_manager: Option<TriggerDdlManagerRef> = plugins.get();
let ddl_task_executor = Self::create_ddl_task_executor(
procedure_manager.clone(),
node_manager.clone(),
layered_cache_registry.clone(),
table_metadata_manager,
table_meta_allocator,
flow_metadata_manager,
flow_meta_allocator,
#[cfg(feature = "enterprise")]
trigger_ddl_manager,
)
.await?;
let ddl_manager = {
let trigger_ddl_manager: Option<common_meta::ddl_manager::TriggerDdlManagerRef> =
plugins.get();
ddl_manager.with_trigger_ddl_manager(trigger_ddl_manager)
};
let ddl_task_executor: ProcedureExecutorRef = Arc::new(ddl_manager);
let fe_instance = FrontendBuilder::new(
fe_opts.clone(),
@@ -658,7 +689,7 @@ impl StartCommand {
let export_metrics_task = ExportMetricsTask::try_new(&opts.export_metrics, Some(&plugins))
.context(error::ServersSnafu)?;
let servers = Services::new(opts, fe_instance.clone(), plugins)
let servers = Services::new(opts, fe_instance.clone(), plugins.clone())
.build()
.context(error::StartFrontendSnafu)?;
@@ -669,51 +700,26 @@ impl StartCommand {
export_metrics_task,
};
#[cfg(feature = "enterprise")]
let components = Components {
plugins,
kv_backend,
frontend_client,
catalog_manager,
};
Ok(Instance {
datanode,
frontend,
flownode,
procedure_manager,
wal_options_allocator,
#[cfg(feature = "enterprise")]
components,
_guard: guard,
})
}
#[allow(clippy::too_many_arguments)]
pub async fn create_ddl_task_executor(
procedure_manager: ProcedureManagerRef,
node_manager: NodeManagerRef,
cache_invalidator: CacheInvalidatorRef,
table_metadata_manager: TableMetadataManagerRef,
table_metadata_allocator: TableMetadataAllocatorRef,
flow_metadata_manager: FlowMetadataManagerRef,
flow_metadata_allocator: FlowMetadataAllocatorRef,
#[cfg(feature = "enterprise")] trigger_ddl_manager: Option<TriggerDdlManagerRef>,
) -> Result<ProcedureExecutorRef> {
let procedure_executor: ProcedureExecutorRef = Arc::new(
DdlManager::try_new(
DdlContext {
node_manager,
cache_invalidator,
memory_region_keeper: Arc::new(MemoryRegionKeeper::default()),
leader_region_registry: Arc::new(LeaderRegionRegistry::default()),
table_metadata_manager,
table_metadata_allocator,
flow_metadata_manager,
flow_metadata_allocator,
region_failure_detector_controller: Arc::new(NoopRegionFailureDetectorControl),
},
procedure_manager,
true,
#[cfg(feature = "enterprise")]
trigger_ddl_manager,
)
.context(error::InitDdlManagerSnafu)?,
);
Ok(procedure_executor)
}
pub async fn create_table_metadata_manager(
kv_backend: KvBackendRef,
) -> Result<TableMetadataManagerRef> {
@@ -849,7 +855,7 @@ mod tests {
use common_config::ENV_VAR_SEP;
use common_test_util::temp_dir::create_named_temp_file;
use common_wal::config::DatanodeWalConfig;
use datanode::config::{FileConfig, GcsConfig};
use object_store::config::{FileConfig, GcsConfig};
use super::*;
use crate::options::GlobalOptions;
@@ -968,15 +974,15 @@ mod tests {
assert!(matches!(
&dn_opts.storage.store,
datanode::config::ObjectStoreConfig::File(FileConfig { .. })
object_store::config::ObjectStoreConfig::File(FileConfig { .. })
));
assert_eq!(dn_opts.storage.providers.len(), 2);
assert!(matches!(
dn_opts.storage.providers[0],
datanode::config::ObjectStoreConfig::Gcs(GcsConfig { .. })
object_store::config::ObjectStoreConfig::Gcs(GcsConfig { .. })
));
match &dn_opts.storage.providers[1] {
datanode::config::ObjectStoreConfig::S3(s3_config) => {
object_store::config::ObjectStoreConfig::S3(s3_config) => {
assert_eq!(
"SecretBox<alloc::string::String>([REDACTED])".to_string(),
format!("{:?}", s3_config.access_key_id)

View File

@@ -18,17 +18,19 @@ use cmd::options::GreptimeOptions;
use cmd::standalone::StandaloneOptions;
use common_config::{Configurable, DEFAULT_DATA_HOME};
use common_options::datanode::{ClientOptions, DatanodeClientOptions};
use common_telemetry::logging::{LoggingOptions, DEFAULT_LOGGING_DIR, DEFAULT_OTLP_ENDPOINT};
use common_telemetry::logging::{LoggingOptions, DEFAULT_LOGGING_DIR, DEFAULT_OTLP_HTTP_ENDPOINT};
use common_wal::config::raft_engine::RaftEngineConfig;
use common_wal::config::DatanodeWalConfig;
use datanode::config::{DatanodeOptions, RegionEngineConfig, StorageConfig};
use file_engine::config::EngineConfig as FileEngineConfig;
use flow::FlownodeOptions;
use frontend::frontend::FrontendOptions;
use meta_client::MetaClientOptions;
use meta_srv::metasrv::MetasrvOptions;
use meta_srv::selector::SelectorType;
use metric_engine::config::EngineConfig as MetricEngineConfig;
use mito2::config::MitoConfig;
use query::options::QueryOptions;
use servers::export_metrics::ExportMetricsOption;
use servers::grpc::GrpcOptions;
use servers::http::HttpOptions;
@@ -81,7 +83,7 @@ fn test_load_datanode_example_config() {
logging: LoggingOptions {
level: Some("info".to_string()),
dir: format!("{}/{}", DEFAULT_DATA_HOME, DEFAULT_LOGGING_DIR),
otlp_endpoint: Some(DEFAULT_OTLP_ENDPOINT.to_string()),
otlp_endpoint: Some(DEFAULT_OTLP_HTTP_ENDPOINT.to_string()),
tracing_sample_ratio: Some(Default::default()),
..Default::default()
},
@@ -124,7 +126,7 @@ fn test_load_frontend_example_config() {
logging: LoggingOptions {
level: Some("info".to_string()),
dir: format!("{}/{}", DEFAULT_DATA_HOME, DEFAULT_LOGGING_DIR),
otlp_endpoint: Some(DEFAULT_OTLP_ENDPOINT.to_string()),
otlp_endpoint: Some(DEFAULT_OTLP_HTTP_ENDPOINT.to_string()),
tracing_sample_ratio: Some(Default::default()),
..Default::default()
},
@@ -172,7 +174,7 @@ fn test_load_metasrv_example_config() {
logging: LoggingOptions {
dir: format!("{}/{}", DEFAULT_DATA_HOME, DEFAULT_LOGGING_DIR),
level: Some("info".to_string()),
otlp_endpoint: Some(DEFAULT_OTLP_ENDPOINT.to_string()),
otlp_endpoint: Some(DEFAULT_OTLP_HTTP_ENDPOINT.to_string()),
tracing_sample_ratio: Some(Default::default()),
..Default::default()
},
@@ -195,6 +197,57 @@ fn test_load_metasrv_example_config() {
similar_asserts::assert_eq!(options, expected);
}
#[test]
fn test_load_flownode_example_config() {
let example_config = common_test_util::find_workspace_path("config/flownode.example.toml");
let options =
GreptimeOptions::<FlownodeOptions>::load_layered_options(example_config.to_str(), "")
.unwrap();
let expected = GreptimeOptions::<FlownodeOptions> {
component: FlownodeOptions {
node_id: Some(14),
flow: Default::default(),
grpc: GrpcOptions {
bind_addr: "127.0.0.1:6800".to_string(),
server_addr: "127.0.0.1:6800".to_string(),
runtime_size: 2,
..Default::default()
},
logging: LoggingOptions {
dir: format!("{}/{}", DEFAULT_DATA_HOME, DEFAULT_LOGGING_DIR),
level: Some("info".to_string()),
otlp_endpoint: Some(DEFAULT_OTLP_HTTP_ENDPOINT.to_string()),
otlp_export_protocol: Some(common_telemetry::logging::OtlpExportProtocol::Http),
tracing_sample_ratio: Some(Default::default()),
..Default::default()
},
tracing: Default::default(),
heartbeat: Default::default(),
// flownode deliberately use a slower query parallelism
// to avoid overwhelming the frontend with too many queries
query: QueryOptions { parallelism: 1 },
meta_client: Some(MetaClientOptions {
metasrv_addrs: vec!["127.0.0.1:3002".to_string()],
timeout: Duration::from_secs(3),
heartbeat_timeout: Duration::from_millis(500),
ddl_timeout: Duration::from_secs(10),
connect_timeout: Duration::from_secs(1),
tcp_nodelay: true,
metadata_cache_max_capacity: 100000,
metadata_cache_ttl: Duration::from_secs(600),
metadata_cache_tti: Duration::from_secs(300),
}),
http: HttpOptions {
addr: "127.0.0.1:4000".to_string(),
..Default::default()
},
user_provider: None,
},
..Default::default()
};
similar_asserts::assert_eq!(options, expected);
}
#[test]
fn test_load_standalone_example_config() {
let example_config = common_test_util::find_workspace_path("config/standalone.example.toml");
@@ -229,7 +282,7 @@ fn test_load_standalone_example_config() {
logging: LoggingOptions {
level: Some("info".to_string()),
dir: format!("{}/{}", DEFAULT_DATA_HOME, DEFAULT_LOGGING_DIR),
otlp_endpoint: Some(DEFAULT_OTLP_ENDPOINT.to_string()),
otlp_endpoint: Some(DEFAULT_OTLP_HTTP_ENDPOINT.to_string()),
tracing_sample_ratio: Some(Default::default()),
..Default::default()
},

View File

@@ -78,7 +78,7 @@ pub const INFORMATION_SCHEMA_ROUTINES_TABLE_ID: u32 = 21;
pub const INFORMATION_SCHEMA_SCHEMA_PRIVILEGES_TABLE_ID: u32 = 22;
/// id for information_schema.TABLE_PRIVILEGES
pub const INFORMATION_SCHEMA_TABLE_PRIVILEGES_TABLE_ID: u32 = 23;
/// id for information_schema.TRIGGERS
/// id for information_schema.TRIGGERS (for mysql)
pub const INFORMATION_SCHEMA_TRIGGERS_TABLE_ID: u32 = 24;
/// id for information_schema.GLOBAL_STATUS
pub const INFORMATION_SCHEMA_GLOBAL_STATUS_TABLE_ID: u32 = 25;
@@ -104,6 +104,8 @@ pub const INFORMATION_SCHEMA_PROCEDURE_INFO_TABLE_ID: u32 = 34;
pub const INFORMATION_SCHEMA_REGION_STATISTICS_TABLE_ID: u32 = 35;
/// id for information_schema.process_list
pub const INFORMATION_SCHEMA_PROCESS_LIST_TABLE_ID: u32 = 36;
/// id for information_schema.trigger_list (for greptimedb trigger)
pub const INFORMATION_SCHEMA_TRIGGER_TABLE_ID: u32 = 37;
// ----- End of information_schema tables -----

View File

@@ -14,6 +14,7 @@ common-macro.workspace = true
config.workspace = true
humantime-serde.workspace = true
num_cpus.workspace = true
object-store.workspace = true
serde.workspace = true
serde_json.workspace = true
serde_with.workspace = true

View File

@@ -106,7 +106,7 @@ mod tests {
use common_telemetry::logging::LoggingOptions;
use common_test_util::temp_dir::create_named_temp_file;
use common_wal::config::DatanodeWalConfig;
use datanode::config::{ObjectStoreConfig, StorageConfig};
use datanode::config::StorageConfig;
use meta_client::MetaClientOptions;
use serde::{Deserialize, Serialize};
@@ -212,7 +212,7 @@ mod tests {
// Check the configs from environment variables.
match &opts.storage.store {
ObjectStoreConfig::S3(s3_config) => {
object_store::config::ObjectStoreConfig::S3(s3_config) => {
assert_eq!(s3_config.bucket, "mybucket".to_string());
}
_ => panic!("unexpected store type"),

View File

@@ -119,6 +119,11 @@ pub enum StatusCode {
FlowAlreadyExists = 8000,
FlowNotFound = 8001,
// ====== End of flow related status code =====
// ====== Begin of trigger related status code =====
TriggerAlreadyExists = 9000,
TriggerNotFound = 9001,
// ====== End of trigger related status code =====
}
impl StatusCode {
@@ -155,6 +160,8 @@ impl StatusCode {
| StatusCode::RegionNotFound
| StatusCode::FlowAlreadyExists
| StatusCode::FlowNotFound
| StatusCode::TriggerAlreadyExists
| StatusCode::TriggerNotFound
| StatusCode::RegionReadonly
| StatusCode::TableColumnNotFound
| StatusCode::TableColumnExists
@@ -198,6 +205,8 @@ impl StatusCode {
| StatusCode::PlanQuery
| StatusCode::FlowAlreadyExists
| StatusCode::FlowNotFound
| StatusCode::TriggerAlreadyExists
| StatusCode::TriggerNotFound
| StatusCode::RegionNotReady
| StatusCode::RegionBusy
| StatusCode::RegionReadonly
@@ -230,6 +239,48 @@ impl fmt::Display for StatusCode {
}
}
#[macro_export]
macro_rules! define_from_tonic_status {
($Error: ty, $Variant: ident) => {
impl From<tonic::Status> for $Error {
fn from(e: tonic::Status) -> Self {
use snafu::location;
fn metadata_value(e: &tonic::Status, key: &str) -> Option<String> {
e.metadata()
.get(key)
.and_then(|v| String::from_utf8(v.as_bytes().to_vec()).ok())
}
let code = metadata_value(&e, $crate::GREPTIME_DB_HEADER_ERROR_CODE)
.and_then(|s| {
if let Ok(code) = s.parse::<u32>() {
StatusCode::from_u32(code)
} else {
None
}
})
.unwrap_or_else(|| match e.code() {
tonic::Code::Cancelled => StatusCode::Cancelled,
tonic::Code::DeadlineExceeded => StatusCode::DeadlineExceeded,
_ => StatusCode::Internal,
});
let msg = metadata_value(&e, $crate::GREPTIME_DB_HEADER_ERROR_MSG)
.unwrap_or_else(|| e.message().to_string());
// TODO(LFC): Make the error variant defined automatically.
Self::$Variant {
code,
msg,
tonic_code: e.code(),
location: location!(),
}
}
}
};
}
#[macro_export]
macro_rules! define_into_tonic_status {
($Error: ty) => {
@@ -281,12 +332,14 @@ pub fn status_to_tonic_code(status_code: StatusCode) -> Code {
| StatusCode::TableColumnExists
| StatusCode::RegionAlreadyExists
| StatusCode::DatabaseAlreadyExists
| StatusCode::TriggerAlreadyExists
| StatusCode::FlowAlreadyExists => Code::AlreadyExists,
StatusCode::TableNotFound
| StatusCode::RegionNotFound
| StatusCode::TableColumnNotFound
| StatusCode::DatabaseNotFound
| StatusCode::UserNotFound
| StatusCode::TriggerNotFound
| StatusCode::FlowNotFound => Code::NotFound,
StatusCode::TableUnavailable
| StatusCode::StorageUnavailable
@@ -304,15 +357,6 @@ pub fn status_to_tonic_code(status_code: StatusCode) -> Code {
}
}
/// Converts tonic [Code] to [StatusCode].
pub fn convert_tonic_code_to_status_code(code: Code) -> StatusCode {
match code {
Code::Cancelled => StatusCode::Cancelled,
Code::DeadlineExceeded => StatusCode::DeadlineExceeded,
_ => StatusCode::Internal,
}
}
#[cfg(test)]
mod tests {
use strum::IntoEnumIterator;

View File

@@ -12,6 +12,7 @@
// See the License for the specific language governing permissions and
// limitations under the License.
use std::fmt::Debug;
use std::time::Duration;
use common_grpc::channel_manager::{ChannelConfig, ChannelManager};
@@ -30,7 +31,7 @@ use crate::error::{MetaSnafu, Result};
pub type FrontendClientPtr = Box<dyn FrontendClient>;
#[async_trait::async_trait]
pub trait FrontendClient: Send {
pub trait FrontendClient: Send + Debug {
async fn list_process(&mut self, req: ListProcessRequest) -> Result<ListProcessResponse>;
async fn kill_process(&mut self, req: KillProcessRequest) -> Result<KillProcessResponse>;

View File

@@ -33,6 +33,7 @@ common-version.workspace = true
datafusion.workspace = true
datafusion-common.workspace = true
datafusion-expr.workspace = true
datafusion-functions-aggregate-common.workspace = true
datatypes.workspace = true
derive_more = { version = "1", default-features = false, features = ["display"] }
geo = { version = "0.29", optional = true }

View File

@@ -13,6 +13,7 @@
// limitations under the License.
pub mod approximate;
pub mod count_hash;
#[cfg(feature = "geo")]
pub mod geo;
pub mod vector;

View File

@@ -0,0 +1,647 @@
// 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.
//! `CountHash` / `count_hash` is a hash-based approximate distinct count function.
//!
//! It is a variant of `CountDistinct` that uses a hash function to approximate the
//! distinct count.
//! It is designed to be more efficient than `CountDistinct` for large datasets,
//! but it is not as accurate, as the hash value may be collision.
use std::collections::HashSet;
use std::fmt::Debug;
use std::sync::Arc;
use ahash::RandomState;
use datafusion_common::cast::as_list_array;
use datafusion_common::error::Result;
use datafusion_common::hash_utils::create_hashes;
use datafusion_common::utils::SingleRowListArrayBuilder;
use datafusion_common::{internal_err, not_impl_err, ScalarValue};
use datafusion_expr::function::{AccumulatorArgs, StateFieldsArgs};
use datafusion_expr::utils::{format_state_name, AggregateOrderSensitivity};
use datafusion_expr::{
Accumulator, AggregateUDF, AggregateUDFImpl, EmitTo, GroupsAccumulator, ReversedUDAF,
SetMonotonicity, Signature, TypeSignature, Volatility,
};
use datafusion_functions_aggregate_common::aggregate::groups_accumulator::nulls::filtered_null_mask;
use datatypes::arrow;
use datatypes::arrow::array::{
Array, ArrayRef, AsArray, BooleanArray, Int64Array, ListArray, UInt64Array,
};
use datatypes::arrow::buffer::{OffsetBuffer, ScalarBuffer};
use datatypes::arrow::datatypes::{DataType, Field};
use crate::function_registry::FunctionRegistry;
type HashValueType = u64;
// read from /dev/urandom 4047821dc6144e4b2abddf23ad4171126a52eeecd26eff2191cf673b965a7875
const RANDOM_SEED_0: u64 = 0x4047821dc6144e4b;
const RANDOM_SEED_1: u64 = 0x2abddf23ad417112;
const RANDOM_SEED_2: u64 = 0x6a52eeecd26eff21;
const RANDOM_SEED_3: u64 = 0x91cf673b965a7875;
impl CountHash {
pub fn register(registry: &FunctionRegistry) {
registry.register_aggr(CountHash::udf_impl());
}
pub fn udf_impl() -> AggregateUDF {
AggregateUDF::new_from_impl(CountHash {
signature: Signature::one_of(
vec![TypeSignature::VariadicAny, TypeSignature::Nullary],
Volatility::Immutable,
),
})
}
}
#[derive(Debug, Clone)]
pub struct CountHash {
signature: Signature,
}
impl AggregateUDFImpl for CountHash {
fn as_any(&self) -> &dyn std::any::Any {
self
}
fn name(&self) -> &str {
"count_hash"
}
fn signature(&self) -> &Signature {
&self.signature
}
fn return_type(&self, _arg_types: &[DataType]) -> Result<DataType> {
Ok(DataType::Int64)
}
fn is_nullable(&self) -> bool {
false
}
fn state_fields(&self, args: StateFieldsArgs) -> Result<Vec<Field>> {
Ok(vec![Field::new_list(
format_state_name(args.name, "count_hash"),
Field::new_list_field(DataType::UInt64, true),
// For count_hash accumulator, null list item stands for an
// empty value set (i.e., all NULL value so far for that group).
true,
)])
}
fn accumulator(&self, acc_args: AccumulatorArgs) -> Result<Box<dyn Accumulator>> {
if acc_args.exprs.len() > 1 {
return not_impl_err!("count_hash with multiple arguments");
}
Ok(Box::new(CountHashAccumulator {
values: HashSet::default(),
random_state: RandomState::with_seeds(
RANDOM_SEED_0,
RANDOM_SEED_1,
RANDOM_SEED_2,
RANDOM_SEED_3,
),
batch_hashes: vec![],
}))
}
fn aliases(&self) -> &[String] {
&[]
}
fn groups_accumulator_supported(&self, _args: AccumulatorArgs) -> bool {
true
}
fn create_groups_accumulator(
&self,
args: AccumulatorArgs,
) -> Result<Box<dyn GroupsAccumulator>> {
if args.exprs.len() > 1 {
return not_impl_err!("count_hash with multiple arguments");
}
Ok(Box::new(CountHashGroupAccumulator::new()))
}
fn reverse_expr(&self) -> ReversedUDAF {
ReversedUDAF::Identical
}
fn order_sensitivity(&self) -> AggregateOrderSensitivity {
AggregateOrderSensitivity::Insensitive
}
fn default_value(&self, _data_type: &DataType) -> Result<ScalarValue> {
Ok(ScalarValue::Int64(Some(0)))
}
fn set_monotonicity(&self, _data_type: &DataType) -> SetMonotonicity {
SetMonotonicity::Increasing
}
}
/// GroupsAccumulator for `count_hash` aggregate function
#[derive(Debug)]
pub struct CountHashGroupAccumulator {
/// One HashSet per group to track distinct values
distinct_sets: Vec<HashSet<HashValueType, RandomState>>,
random_state: RandomState,
batch_hashes: Vec<HashValueType>,
}
impl Default for CountHashGroupAccumulator {
fn default() -> Self {
Self::new()
}
}
impl CountHashGroupAccumulator {
pub fn new() -> Self {
Self {
distinct_sets: vec![],
random_state: RandomState::with_seeds(
RANDOM_SEED_0,
RANDOM_SEED_1,
RANDOM_SEED_2,
RANDOM_SEED_3,
),
batch_hashes: vec![],
}
}
fn ensure_sets(&mut self, total_num_groups: usize) {
if self.distinct_sets.len() < total_num_groups {
self.distinct_sets
.resize_with(total_num_groups, HashSet::default);
}
}
}
impl GroupsAccumulator for CountHashGroupAccumulator {
fn update_batch(
&mut self,
values: &[ArrayRef],
group_indices: &[usize],
opt_filter: Option<&BooleanArray>,
total_num_groups: usize,
) -> Result<()> {
assert_eq!(values.len(), 1, "count_hash expects a single argument");
self.ensure_sets(total_num_groups);
let array = &values[0];
self.batch_hashes.clear();
self.batch_hashes.resize(array.len(), 0);
let hashes = create_hashes(
&[ArrayRef::clone(array)],
&self.random_state,
&mut self.batch_hashes,
)?;
// Use a pattern similar to accumulate_indices to process rows
// that are not null and pass the filter
let nulls = array.logical_nulls();
match (nulls.as_ref(), opt_filter) {
(None, None) => {
// No nulls, no filter - process all rows
for (row_idx, &group_idx) in group_indices.iter().enumerate() {
self.distinct_sets[group_idx].insert(hashes[row_idx]);
}
}
(Some(nulls), None) => {
// Has nulls, no filter
for (row_idx, (&group_idx, is_valid)) in
group_indices.iter().zip(nulls.iter()).enumerate()
{
if is_valid {
self.distinct_sets[group_idx].insert(hashes[row_idx]);
}
}
}
(None, Some(filter)) => {
// No nulls, has filter
for (row_idx, (&group_idx, filter_value)) in
group_indices.iter().zip(filter.iter()).enumerate()
{
if let Some(true) = filter_value {
self.distinct_sets[group_idx].insert(hashes[row_idx]);
}
}
}
(Some(nulls), Some(filter)) => {
// Has nulls and filter
let iter = filter
.iter()
.zip(group_indices.iter())
.zip(nulls.iter())
.enumerate();
for (row_idx, ((filter_value, &group_idx), is_valid)) in iter {
if is_valid && filter_value == Some(true) {
self.distinct_sets[group_idx].insert(hashes[row_idx]);
}
}
}
}
Ok(())
}
fn evaluate(&mut self, emit_to: EmitTo) -> Result<ArrayRef> {
let distinct_sets: Vec<HashSet<u64, RandomState>> =
emit_to.take_needed(&mut self.distinct_sets);
let counts = distinct_sets
.iter()
.map(|set| set.len() as i64)
.collect::<Vec<_>>();
Ok(Arc::new(Int64Array::from(counts)))
}
fn merge_batch(
&mut self,
values: &[ArrayRef],
group_indices: &[usize],
_opt_filter: Option<&BooleanArray>,
total_num_groups: usize,
) -> Result<()> {
assert_eq!(
values.len(),
1,
"count_hash merge expects a single state array"
);
self.ensure_sets(total_num_groups);
let list_array = as_list_array(&values[0])?;
// For each group in the incoming batch
for (i, &group_idx) in group_indices.iter().enumerate() {
if i < list_array.len() {
let inner_array = list_array.value(i);
let inner_array = inner_array.as_any().downcast_ref::<UInt64Array>().unwrap();
// Add each value to our set for this group
for j in 0..inner_array.len() {
if !inner_array.is_null(j) {
self.distinct_sets[group_idx].insert(inner_array.value(j));
}
}
}
}
Ok(())
}
fn state(&mut self, emit_to: EmitTo) -> Result<Vec<ArrayRef>> {
let distinct_sets: Vec<HashSet<u64, RandomState>> =
emit_to.take_needed(&mut self.distinct_sets);
let mut offsets = Vec::with_capacity(distinct_sets.len() + 1);
offsets.push(0);
let mut curr_len = 0i32;
let mut value_iter = distinct_sets
.into_iter()
.flat_map(|set| {
// build offset
curr_len += set.len() as i32;
offsets.push(curr_len);
// convert into iter
set.into_iter()
})
.peekable();
let data_array: ArrayRef = if value_iter.peek().is_none() {
arrow::array::new_empty_array(&DataType::UInt64) as _
} else {
Arc::new(UInt64Array::from_iter_values(value_iter))
};
let offset_buffer = OffsetBuffer::new(ScalarBuffer::from(offsets));
let list_array = ListArray::new(
Arc::new(Field::new_list_field(DataType::UInt64, true)),
offset_buffer,
data_array,
None,
);
Ok(vec![Arc::new(list_array) as _])
}
fn convert_to_state(
&self,
values: &[ArrayRef],
opt_filter: Option<&BooleanArray>,
) -> Result<Vec<ArrayRef>> {
// For a single hash value per row, create a list array with that value
assert_eq!(values.len(), 1, "count_hash expects a single argument");
let values = ArrayRef::clone(&values[0]);
let offsets = OffsetBuffer::new(ScalarBuffer::from_iter(0..values.len() as i32 + 1));
let nulls = filtered_null_mask(opt_filter, &values);
let list_array = ListArray::new(
Arc::new(Field::new_list_field(DataType::UInt64, true)),
offsets,
values,
nulls,
);
Ok(vec![Arc::new(list_array)])
}
fn supports_convert_to_state(&self) -> bool {
true
}
fn size(&self) -> usize {
// Base size of the struct
let mut size = size_of::<Self>();
// Size of the vector holding the HashSets
size += size_of::<Vec<HashSet<HashValueType, RandomState>>>()
+ self.distinct_sets.capacity() * size_of::<HashSet<HashValueType, RandomState>>();
// Estimate HashSet contents size more efficiently
// Instead of iterating through all values which is expensive, use an approximation
for set in &self.distinct_sets {
// Base size of the HashSet
size += set.capacity() * size_of::<HashValueType>();
}
size
}
}
#[derive(Debug)]
struct CountHashAccumulator {
values: HashSet<HashValueType, RandomState>,
random_state: RandomState,
batch_hashes: Vec<HashValueType>,
}
impl CountHashAccumulator {
// calculating the size for fixed length values, taking first batch size *
// number of batches.
fn fixed_size(&self) -> usize {
size_of_val(self) + (size_of::<HashValueType>() * self.values.capacity())
}
}
impl Accumulator for CountHashAccumulator {
/// Returns the distinct values seen so far as (one element) ListArray.
fn state(&mut self) -> Result<Vec<ScalarValue>> {
let values = self.values.iter().cloned().collect::<Vec<_>>();
let arr = Arc::new(UInt64Array::from(values)) as _;
let list_scalar = SingleRowListArrayBuilder::new(arr).build_list_scalar();
Ok(vec![list_scalar])
}
fn update_batch(&mut self, values: &[ArrayRef]) -> Result<()> {
if values.is_empty() {
return Ok(());
}
let arr = &values[0];
if arr.data_type() == &DataType::Null {
return Ok(());
}
self.batch_hashes.clear();
self.batch_hashes.resize(arr.len(), 0);
let hashes = create_hashes(
&[ArrayRef::clone(arr)],
&self.random_state,
&mut self.batch_hashes,
)?;
for hash in hashes.as_slice() {
self.values.insert(*hash);
}
Ok(())
}
/// Merges multiple sets of distinct values into the current set.
///
/// The input to this function is a `ListArray` with **multiple** rows,
/// where each row contains the values from a partial aggregate's phase (e.g.
/// the result of calling `Self::state` on multiple accumulators).
fn merge_batch(&mut self, states: &[ArrayRef]) -> Result<()> {
if states.is_empty() {
return Ok(());
}
assert_eq!(states.len(), 1, "array_agg states must be singleton!");
let array = &states[0];
let list_array = array.as_list::<i32>();
for inner_array in list_array.iter() {
let Some(inner_array) = inner_array else {
return internal_err!(
"Intermediate results of count_hash should always be non null"
);
};
let hash_array = inner_array.as_any().downcast_ref::<UInt64Array>().unwrap();
for i in 0..hash_array.len() {
self.values.insert(hash_array.value(i));
}
}
Ok(())
}
fn evaluate(&mut self) -> Result<ScalarValue> {
Ok(ScalarValue::Int64(Some(self.values.len() as i64)))
}
fn size(&self) -> usize {
self.fixed_size()
}
}
#[cfg(test)]
mod tests {
use datatypes::arrow::array::{Array, BooleanArray, Int32Array, Int64Array};
use super::*;
fn create_test_accumulator() -> CountHashAccumulator {
CountHashAccumulator {
values: HashSet::default(),
random_state: RandomState::with_seeds(
RANDOM_SEED_0,
RANDOM_SEED_1,
RANDOM_SEED_2,
RANDOM_SEED_3,
),
batch_hashes: vec![],
}
}
#[test]
fn test_count_hash_accumulator() -> Result<()> {
let mut acc = create_test_accumulator();
// Test with some data
let array = Arc::new(Int32Array::from(vec![
Some(1),
Some(2),
Some(3),
Some(1),
Some(2),
None,
])) as ArrayRef;
acc.update_batch(&[array])?;
let result = acc.evaluate()?;
assert_eq!(result, ScalarValue::Int64(Some(4)));
// Test with empty data
let mut acc = create_test_accumulator();
let array = Arc::new(Int32Array::from(vec![] as Vec<Option<i32>>)) as ArrayRef;
acc.update_batch(&[array])?;
let result = acc.evaluate()?;
assert_eq!(result, ScalarValue::Int64(Some(0)));
// Test with only nulls
let mut acc = create_test_accumulator();
let array = Arc::new(Int32Array::from(vec![None, None, None])) as ArrayRef;
acc.update_batch(&[array])?;
let result = acc.evaluate()?;
assert_eq!(result, ScalarValue::Int64(Some(1)));
Ok(())
}
#[test]
fn test_count_hash_accumulator_merge() -> Result<()> {
// Accumulator 1
let mut acc1 = create_test_accumulator();
let array1 = Arc::new(Int32Array::from(vec![Some(1), Some(2), Some(3)])) as ArrayRef;
acc1.update_batch(&[array1])?;
let state1 = acc1.state()?;
// Accumulator 2
let mut acc2 = create_test_accumulator();
let array2 = Arc::new(Int32Array::from(vec![Some(3), Some(4), Some(5)])) as ArrayRef;
acc2.update_batch(&[array2])?;
let state2 = acc2.state()?;
// Merge state1 and state2 into a new accumulator
let mut acc_merged = create_test_accumulator();
let state_array1 = state1[0].to_array()?;
let state_array2 = state2[0].to_array()?;
acc_merged.merge_batch(&[state_array1])?;
acc_merged.merge_batch(&[state_array2])?;
let result = acc_merged.evaluate()?;
// Distinct values are {1, 2, 3, 4, 5}, so count is 5
assert_eq!(result, ScalarValue::Int64(Some(5)));
Ok(())
}
fn create_test_group_accumulator() -> CountHashGroupAccumulator {
CountHashGroupAccumulator::new()
}
#[test]
fn test_count_hash_group_accumulator() -> Result<()> {
let mut acc = create_test_group_accumulator();
let values = Arc::new(Int32Array::from(vec![1, 2, 1, 3, 2, 4, 5])) as ArrayRef;
let group_indices = vec![0, 1, 0, 0, 1, 2, 0];
let total_num_groups = 3;
acc.update_batch(&[values], &group_indices, None, total_num_groups)?;
let result_array = acc.evaluate(EmitTo::All)?;
let result = result_array.as_any().downcast_ref::<Int64Array>().unwrap();
// Group 0: {1, 3, 5} -> 3
// Group 1: {2} -> 1
// Group 2: {4} -> 1
assert_eq!(result.value(0), 3);
assert_eq!(result.value(1), 1);
assert_eq!(result.value(2), 1);
Ok(())
}
#[test]
fn test_count_hash_group_accumulator_with_filter() -> Result<()> {
let mut acc = create_test_group_accumulator();
let values = Arc::new(Int32Array::from(vec![1, 2, 3, 4, 5, 6])) as ArrayRef;
let group_indices = vec![0, 0, 1, 1, 2, 2];
let filter = BooleanArray::from(vec![true, false, true, true, false, true]);
let total_num_groups = 3;
acc.update_batch(&[values], &group_indices, Some(&filter), total_num_groups)?;
let result_array = acc.evaluate(EmitTo::All)?;
let result = result_array.as_any().downcast_ref::<Int64Array>().unwrap();
// Group 0: {1} (2 is filtered out) -> 1
// Group 1: {3, 4} -> 2
// Group 2: {6} (5 is filtered out) -> 1
assert_eq!(result.value(0), 1);
assert_eq!(result.value(1), 2);
assert_eq!(result.value(2), 1);
Ok(())
}
#[test]
fn test_count_hash_group_accumulator_merge() -> Result<()> {
// Accumulator 1
let mut acc1 = create_test_group_accumulator();
let values1 = Arc::new(Int32Array::from(vec![1, 2, 3, 4])) as ArrayRef;
let group_indices1 = vec![0, 0, 1, 1];
acc1.update_batch(&[values1], &group_indices1, None, 2)?;
// acc1 state: group 0 -> {1, 2}, group 1 -> {3, 4}
let state1 = acc1.state(EmitTo::All)?;
// Accumulator 2
let mut acc2 = create_test_group_accumulator();
let values2 = Arc::new(Int32Array::from(vec![5, 6, 1, 3])) as ArrayRef;
// Merge into different group indices
let group_indices2 = vec![2, 2, 0, 1];
acc2.update_batch(&[values2], &group_indices2, None, 3)?;
// acc2 state: group 0 -> {1}, group 1 -> {3}, group 2 -> {5, 6}
// Merge state from acc1 into acc2
// We will merge acc1's group 0 into acc2's group 0
// and acc1's group 1 into acc2's group 2
let merge_group_indices = vec![0, 2];
acc2.merge_batch(&state1, &merge_group_indices, None, 3)?;
let result_array = acc2.evaluate(EmitTo::All)?;
let result = result_array.as_any().downcast_ref::<Int64Array>().unwrap();
// Final state of acc2:
// Group 0: {1} U {1, 2} -> {1, 2}, count = 2
// Group 1: {3}, count = 1
// Group 2: {5, 6} U {3, 4} -> {3, 4, 5, 6}, count = 4
assert_eq!(result.value(0), 2);
assert_eq!(result.value(1), 1);
assert_eq!(result.value(2), 4);
Ok(())
}
#[test]
fn test_size() {
let acc = create_test_group_accumulator();
// Just test it doesn't crash and returns a value.
assert!(acc.size() > 0);
}
}

View File

@@ -21,6 +21,7 @@ use once_cell::sync::Lazy;
use crate::admin::AdminFunction;
use crate::aggrs::approximate::ApproximateFunction;
use crate::aggrs::count_hash::CountHash;
use crate::aggrs::vector::VectorFunction as VectorAggrFunction;
use crate::function::{AsyncFunctionRef, Function, FunctionRef};
use crate::function_factory::ScalarFunctionFactory;
@@ -144,6 +145,9 @@ pub static FUNCTION_REGISTRY: Lazy<Arc<FunctionRegistry>> = Lazy::new(|| {
// Approximate functions
ApproximateFunction::register(&function_registry);
// CountHash function
CountHash::register(&function_registry);
Arc::new(function_registry)
});

View File

@@ -14,7 +14,6 @@
pub mod clamp;
mod modulo;
mod pow;
mod rate;
use std::fmt;
@@ -26,7 +25,6 @@ use datafusion::error::DataFusionError;
use datafusion::logical_expr::Volatility;
use datatypes::prelude::ConcreteDataType;
use datatypes::vectors::VectorRef;
pub use pow::PowFunction;
pub use rate::RateFunction;
use snafu::ResultExt;
@@ -39,7 +37,6 @@ pub(crate) struct MathFunction;
impl MathFunction {
pub fn register(registry: &FunctionRegistry) {
registry.register_scalar(ModuloFunction);
registry.register_scalar(PowFunction);
registry.register_scalar(RateFunction);
registry.register_scalar(RangeFunction);
registry.register_scalar(ClampFunction);

View File

@@ -1,120 +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::fmt;
use std::sync::Arc;
use common_query::error::Result;
use common_query::prelude::{Signature, Volatility};
use datatypes::data_type::DataType;
use datatypes::prelude::ConcreteDataType;
use datatypes::types::LogicalPrimitiveType;
use datatypes::vectors::VectorRef;
use datatypes::with_match_primitive_type_id;
use num::traits::Pow;
use num_traits::AsPrimitive;
use crate::function::{Function, FunctionContext};
use crate::scalars::expression::{scalar_binary_op, EvalContext};
#[derive(Clone, Debug, Default)]
pub struct PowFunction;
impl Function for PowFunction {
fn name(&self) -> &str {
"pow"
}
fn return_type(&self, _input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
Ok(ConcreteDataType::float64_datatype())
}
fn signature(&self) -> Signature {
Signature::uniform(2, ConcreteDataType::numerics(), Volatility::Immutable)
}
fn eval(&self, _func_ctx: &FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
with_match_primitive_type_id!(columns[0].data_type().logical_type_id(), |$S| {
with_match_primitive_type_id!(columns[1].data_type().logical_type_id(), |$T| {
let col = scalar_binary_op::<<$S as LogicalPrimitiveType>::Native, <$T as LogicalPrimitiveType>::Native, f64, _>(&columns[0], &columns[1], scalar_pow, &mut EvalContext::default())?;
Ok(Arc::new(col))
},{
unreachable!()
})
},{
unreachable!()
})
}
}
#[inline]
fn scalar_pow<S, T>(value: Option<S>, base: Option<T>, _ctx: &mut EvalContext) -> Option<f64>
where
S: AsPrimitive<f64>,
T: AsPrimitive<f64>,
{
match (value, base) {
(Some(value), Some(base)) => Some(value.as_().pow(base.as_())),
_ => None,
}
}
impl fmt::Display for PowFunction {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
write!(f, "POW")
}
}
#[cfg(test)]
mod tests {
use common_query::prelude::TypeSignature;
use datatypes::value::Value;
use datatypes::vectors::{Float32Vector, Int8Vector};
use super::*;
use crate::function::FunctionContext;
#[test]
fn test_pow_function() {
let pow = PowFunction;
assert_eq!("pow", pow.name());
assert_eq!(
ConcreteDataType::float64_datatype(),
pow.return_type(&[]).unwrap()
);
assert!(matches!(pow.signature(),
Signature {
type_signature: TypeSignature::Uniform(2, valid_types),
volatility: Volatility::Immutable
} if valid_types == ConcreteDataType::numerics()
));
let values = vec![1.0, 2.0, 3.0];
let bases = vec![0i8, -1i8, 3i8];
let args: Vec<VectorRef> = vec![
Arc::new(Float32Vector::from_vec(values.clone())),
Arc::new(Int8Vector::from_vec(bases.clone())),
];
let vector = pow.eval(&FunctionContext::default(), &args).unwrap();
assert_eq!(3, vector.len());
for i in 0..3 {
let p: f64 = (values[i] as f64).pow(bases[i] as f64);
assert!(matches!(vector.get(i), Value::Float64(v) if v == p));
}
}
}

View File

@@ -34,7 +34,7 @@ use table::requests::{
};
use crate::error::{
InvalidColumnDefSnafu, InvalidSetFulltextOptionRequestSnafu,
InvalidColumnDefSnafu, InvalidIndexOptionSnafu, InvalidSetFulltextOptionRequestSnafu,
InvalidSetSkippingIndexOptionRequestSnafu, InvalidSetTableOptionRequestSnafu,
InvalidUnsetTableOptionRequestSnafu, MissingAlterIndexOptionSnafu, MissingFieldSnafu,
MissingTimestampColumnSnafu, Result, UnknownLocationTypeSnafu,
@@ -126,18 +126,21 @@ pub fn alter_expr_to_request(table_id: TableId, expr: AlterTableExpr) -> Result<
api::v1::set_index::Options::Fulltext(f) => AlterKind::SetIndex {
options: SetIndexOptions::Fulltext {
column_name: f.column_name.clone(),
options: FulltextOptions {
enable: f.enable,
analyzer: as_fulltext_option_analyzer(
options: FulltextOptions::new(
f.enable,
as_fulltext_option_analyzer(
Analyzer::try_from(f.analyzer)
.context(InvalidSetFulltextOptionRequestSnafu)?,
),
case_sensitive: f.case_sensitive,
backend: as_fulltext_option_backend(
f.case_sensitive,
as_fulltext_option_backend(
PbFulltextBackend::try_from(f.backend)
.context(InvalidSetFulltextOptionRequestSnafu)?,
),
},
f.granularity as u32,
f.false_positive_rate,
)
.context(InvalidIndexOptionSnafu)?,
},
},
api::v1::set_index::Options::Inverted(i) => AlterKind::SetIndex {
@@ -148,13 +151,15 @@ pub fn alter_expr_to_request(table_id: TableId, expr: AlterTableExpr) -> Result<
api::v1::set_index::Options::Skipping(s) => AlterKind::SetIndex {
options: SetIndexOptions::Skipping {
column_name: s.column_name,
options: SkippingIndexOptions {
granularity: s.granularity as u32,
index_type: as_skipping_index_type(
options: SkippingIndexOptions::new(
s.granularity as u32,
s.false_positive_rate,
as_skipping_index_type(
PbSkippingIndexType::try_from(s.skipping_index_type)
.context(InvalidSetSkippingIndexOptionRequestSnafu)?,
),
},
)
.context(InvalidIndexOptionSnafu)?,
},
},
},

View File

@@ -153,6 +153,14 @@ pub enum Error {
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Invalid index option"))]
InvalidIndexOption {
#[snafu(implicit)]
location: Location,
#[snafu(source)]
error: datatypes::error::Error,
},
}
pub type Result<T> = std::result::Result<T, Error>;
@@ -180,7 +188,8 @@ impl ErrorExt for Error {
| Error::InvalidUnsetTableOptionRequest { .. }
| Error::InvalidSetFulltextOptionRequest { .. }
| Error::InvalidSetSkippingIndexOptionRequest { .. }
| Error::MissingAlterIndexOption { .. } => StatusCode::InvalidArguments,
| Error::MissingAlterIndexOption { .. }
| Error::InvalidIndexOption { .. } => StatusCode::InvalidArguments,
}
}

View File

@@ -27,17 +27,18 @@ use common_telemetry::{error, info, warn};
use futures_util::future;
pub use region_request::make_alter_region_request;
use serde::{Deserialize, Serialize};
use snafu::{ensure, ResultExt};
use snafu::ResultExt;
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::ALTER_PHYSICAL_EXTENSION_KEY;
use strum::AsRefStr;
use table::metadata::TableId;
use crate::ddl::utils::{
add_peer_context_if_needed, map_to_procedure_error, sync_follower_regions,
add_peer_context_if_needed, extract_column_metadatas, map_to_procedure_error,
sync_follower_regions,
};
use crate::ddl::DdlContext;
use crate::error::{DecodeJsonSnafu, MetadataCorruptionSnafu, Result};
use crate::error::Result;
use crate::instruction::CacheIdent;
use crate::key::table_info::TableInfoValue;
use crate::key::table_route::PhysicalTableRouteValue;
@@ -137,37 +138,13 @@ impl AlterLogicalTablesProcedure {
.into_iter()
.collect::<Result<Vec<_>>>()?;
// Collects responses from datanodes.
let phy_raw_schemas = results
.iter_mut()
.map(|res| res.extensions.remove(ALTER_PHYSICAL_EXTENSION_KEY))
.collect::<Vec<_>>();
if phy_raw_schemas.is_empty() {
self.submit_sync_region_requests(results, &physical_table_route.region_routes)
.await;
self.data.state = AlterTablesState::UpdateMetadata;
return Ok(Status::executing(true));
}
// Verify all the physical schemas are the same
// Safety: previous check ensures this vec is not empty
let first = phy_raw_schemas.first().unwrap();
ensure!(
phy_raw_schemas.iter().all(|x| x == first),
MetadataCorruptionSnafu {
err_msg: "The physical schemas from datanodes are not the same."
}
);
// Decodes the physical raw schemas
if let Some(phy_raw_schema) = first {
self.data.physical_columns =
ColumnMetadata::decode_list(phy_raw_schema).context(DecodeJsonSnafu)?;
if let Some(column_metadatas) =
extract_column_metadatas(&mut results, ALTER_PHYSICAL_EXTENSION_KEY)?
{
self.data.physical_columns = column_metadatas;
} else {
warn!("altering logical table result doesn't contains extension key `{ALTER_PHYSICAL_EXTENSION_KEY}`,leaving the physical table's schema unchanged");
}
self.submit_sync_region_requests(results, &physical_table_route.region_routes)
.await;
self.data.state = AlterTablesState::UpdateMetadata;
@@ -183,7 +160,7 @@ impl AlterLogicalTablesProcedure {
if let Err(err) = sync_follower_regions(
&self.context,
self.data.physical_table_id,
results,
&results,
region_routes,
table_info.meta.engine.as_str(),
)

View File

@@ -29,19 +29,22 @@ use common_procedure::{
Context as ProcedureContext, ContextProvider, Error as ProcedureError, LockKey, PoisonKey,
PoisonKeys, Procedure, ProcedureId, Status, StringKey,
};
use common_telemetry::{debug, error, info};
use common_telemetry::{debug, error, info, warn};
use futures::future::{self};
use serde::{Deserialize, Serialize};
use snafu::{ensure, ResultExt};
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::TABLE_COLUMN_METADATA_EXTENSION_KEY;
use store_api::storage::RegionId;
use strum::AsRefStr;
use table::metadata::{RawTableInfo, TableId, TableInfo};
use table::table_reference::TableReference;
use crate::cache_invalidator::Context;
use crate::ddl::physical_table_metadata::update_table_info_column_ids;
use crate::ddl::utils::{
add_peer_context_if_needed, handle_multiple_results, map_to_procedure_error,
sync_follower_regions, MultipleResults,
add_peer_context_if_needed, extract_column_metadatas, handle_multiple_results,
map_to_procedure_error, sync_follower_regions, MultipleResults,
};
use crate::ddl::DdlContext;
use crate::error::{AbortProcedureSnafu, NoLeaderSnafu, PutPoisonSnafu, Result, RetryLaterSnafu};
@@ -202,9 +205,9 @@ impl AlterTableProcedure {
})
}
MultipleResults::Ok(results) => {
self.submit_sync_region_requests(results, &physical_table_route.region_routes)
self.submit_sync_region_requests(&results, &physical_table_route.region_routes)
.await;
self.data.state = AlterTableState::UpdateMetadata;
self.handle_alter_region_response(results)?;
Ok(Status::executing_with_clean_poisons(true))
}
MultipleResults::AllNonRetryable(error) => {
@@ -220,9 +223,22 @@ impl AlterTableProcedure {
}
}
fn handle_alter_region_response(&mut self, mut results: Vec<RegionResponse>) -> Result<()> {
self.data.state = AlterTableState::UpdateMetadata;
if let Some(column_metadatas) =
extract_column_metadatas(&mut results, TABLE_COLUMN_METADATA_EXTENSION_KEY)?
{
self.data.column_metadatas = column_metadatas;
} else {
warn!("altering table result doesn't contains extension key `{TABLE_COLUMN_METADATA_EXTENSION_KEY}`,leaving the table's column metadata unchanged");
}
Ok(())
}
async fn submit_sync_region_requests(
&mut self,
results: Vec<RegionResponse>,
results: &[RegionResponse],
region_routes: &[RegionRoute],
) {
// Safety: filled in `prepare` step.
@@ -268,10 +284,14 @@ impl AlterTableProcedure {
self.on_update_metadata_for_rename(new_table_name.to_string(), table_info_value)
.await?;
} else {
let mut raw_table_info = new_info.into();
if !self.data.column_metadatas.is_empty() {
update_table_info_column_ids(&mut raw_table_info, &self.data.column_metadatas);
}
// region distribution is set in submit_alter_region_requests
let region_distribution = self.data.region_distribution.as_ref().unwrap().clone();
self.on_update_metadata_for_alter(
new_info.into(),
raw_table_info,
region_distribution,
table_info_value,
)
@@ -318,6 +338,16 @@ impl AlterTableProcedure {
lock_key
}
#[cfg(test)]
pub(crate) fn data(&self) -> &AlterTableData {
&self.data
}
#[cfg(test)]
pub(crate) fn mut_data(&mut self) -> &mut AlterTableData {
&mut self.data
}
}
#[async_trait]
@@ -380,6 +410,8 @@ pub struct AlterTableData {
state: AlterTableState,
task: AlterTableTask,
table_id: TableId,
#[serde(default)]
column_metadatas: Vec<ColumnMetadata>,
/// Table info value before alteration.
table_info_value: Option<DeserializedValueWithBytes<TableInfoValue>>,
/// Region distribution for table in case we need to update region options.
@@ -392,6 +424,7 @@ impl AlterTableData {
state: AlterTableState::Prepare,
task,
table_id,
column_metadatas: vec![],
table_info_value: None,
region_distribution: None,
}
@@ -410,4 +443,14 @@ impl AlterTableData {
.as_ref()
.map(|value| &value.table_info)
}
#[cfg(test)]
pub(crate) fn column_metadatas(&self) -> &[ColumnMetadata] {
&self.column_metadatas
}
#[cfg(test)]
pub(crate) fn set_column_metadatas(&mut self, column_metadatas: Vec<ColumnMetadata>) {
self.column_metadatas = column_metadatas;
}
}

View File

@@ -167,6 +167,25 @@ impl CreateFlowProcedure {
}
self.collect_source_tables().await?;
// Validate that source and sink tables are not the same
let sink_table_name = &self.data.task.sink_table_name;
if self
.data
.task
.source_table_names
.iter()
.any(|source| source == sink_table_name)
{
return error::UnsupportedSnafu {
operation: format!(
"Creating flow with source and sink table being the same: {}",
sink_table_name
),
}
.fail();
}
if self.data.flow_id.is_none() {
self.allocate_flow_id().await?;
}

View File

@@ -27,7 +27,7 @@ use common_telemetry::{debug, error, warn};
use futures::future;
pub use region_request::create_region_request_builder;
use serde::{Deserialize, Serialize};
use snafu::{ensure, ResultExt};
use snafu::ResultExt;
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::ALTER_PHYSICAL_EXTENSION_KEY;
use store_api::storage::{RegionId, RegionNumber};
@@ -35,10 +35,11 @@ use strum::AsRefStr;
use table::metadata::{RawTableInfo, TableId};
use crate::ddl::utils::{
add_peer_context_if_needed, map_to_procedure_error, sync_follower_regions,
add_peer_context_if_needed, extract_column_metadatas, map_to_procedure_error,
sync_follower_regions,
};
use crate::ddl::DdlContext;
use crate::error::{DecodeJsonSnafu, MetadataCorruptionSnafu, Result};
use crate::error::Result;
use crate::key::table_route::TableRouteValue;
use crate::lock_key::{CatalogLock, SchemaLock, TableLock, TableNameLock};
use crate::metrics;
@@ -166,47 +167,23 @@ impl CreateLogicalTablesProcedure {
.into_iter()
.collect::<Result<Vec<_>>>()?;
// Collects response from datanodes.
let phy_raw_schemas = results
.iter_mut()
.map(|res| res.extensions.remove(ALTER_PHYSICAL_EXTENSION_KEY))
.collect::<Vec<_>>();
if phy_raw_schemas.is_empty() {
self.submit_sync_region_requests(results, region_routes)
.await;
self.data.state = CreateTablesState::CreateMetadata;
return Ok(Status::executing(false));
}
// Verify all the physical schemas are the same
// Safety: previous check ensures this vec is not empty
let first = phy_raw_schemas.first().unwrap();
ensure!(
phy_raw_schemas.iter().all(|x| x == first),
MetadataCorruptionSnafu {
err_msg: "The physical schemas from datanodes are not the same."
}
);
// Decodes the physical raw schemas
if let Some(phy_raw_schemas) = first {
self.data.physical_columns =
ColumnMetadata::decode_list(phy_raw_schemas).context(DecodeJsonSnafu)?;
if let Some(column_metadatas) =
extract_column_metadatas(&mut results, ALTER_PHYSICAL_EXTENSION_KEY)?
{
self.data.physical_columns = column_metadatas;
} else {
warn!("creating logical table result doesn't contains extension key `{ALTER_PHYSICAL_EXTENSION_KEY}`,leaving the physical table's schema unchanged");
}
self.submit_sync_region_requests(results, region_routes)
self.submit_sync_region_requests(&results, region_routes)
.await;
self.data.state = CreateTablesState::CreateMetadata;
Ok(Status::executing(true))
}
async fn submit_sync_region_requests(
&self,
results: Vec<RegionResponse>,
results: &[RegionResponse],
region_routes: &[RegionRoute],
) {
if let Err(err) = sync_follower_regions(

View File

@@ -22,20 +22,23 @@ use common_procedure::error::{
ExternalSnafu, FromJsonSnafu, Result as ProcedureResult, ToJsonSnafu,
};
use common_procedure::{Context as ProcedureContext, LockKey, Procedure, Status};
use common_telemetry::info;
use common_telemetry::tracing_context::TracingContext;
use common_telemetry::{info, warn};
use futures::future::join_all;
use serde::{Deserialize, Serialize};
use snafu::{ensure, OptionExt, ResultExt};
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::TABLE_COLUMN_METADATA_EXTENSION_KEY;
use store_api::storage::{RegionId, RegionNumber};
use strum::AsRefStr;
use table::metadata::{RawTableInfo, TableId};
use table::table_reference::TableReference;
use crate::ddl::create_table_template::{build_template, CreateRequestBuilder};
use crate::ddl::physical_table_metadata::update_table_info_column_ids;
use crate::ddl::utils::{
add_peer_context_if_needed, convert_region_routes_to_detecting_regions, map_to_procedure_error,
region_storage_path,
add_peer_context_if_needed, convert_region_routes_to_detecting_regions,
extract_column_metadatas, map_to_procedure_error, region_storage_path,
};
use crate::ddl::{DdlContext, TableMetadata};
use crate::error::{self, Result};
@@ -243,14 +246,21 @@ impl CreateTableProcedure {
}
}
join_all(create_region_tasks)
self.creator.data.state = CreateTableState::CreateMetadata;
let mut results = join_all(create_region_tasks)
.await
.into_iter()
.collect::<Result<Vec<_>>>()?;
self.creator.data.state = CreateTableState::CreateMetadata;
if let Some(column_metadatas) =
extract_column_metadatas(&mut results, TABLE_COLUMN_METADATA_EXTENSION_KEY)?
{
self.creator.data.column_metadatas = column_metadatas;
} else {
warn!("creating table result doesn't contains extension key `{TABLE_COLUMN_METADATA_EXTENSION_KEY}`,leaving the table's column metadata unchanged");
}
// TODO(weny): Add more tests.
Ok(Status::executing(true))
}
@@ -262,7 +272,10 @@ impl CreateTableProcedure {
let table_id = self.table_id();
let manager = &self.context.table_metadata_manager;
let raw_table_info = self.table_info().clone();
let mut raw_table_info = self.table_info().clone();
if !self.creator.data.column_metadatas.is_empty() {
update_table_info_column_ids(&mut raw_table_info, &self.creator.data.column_metadatas);
}
// Safety: the region_wal_options must be allocated.
let region_wal_options = self.region_wal_options()?.clone();
// Safety: the table_route must be allocated.
@@ -346,6 +359,7 @@ impl TableCreator {
Self {
data: CreateTableData {
state: CreateTableState::Prepare,
column_metadatas: vec![],
task,
table_route: None,
region_wal_options: None,
@@ -407,6 +421,8 @@ pub enum CreateTableState {
pub struct CreateTableData {
pub state: CreateTableState,
pub task: CreateTableTask,
#[serde(default)]
pub column_metadatas: Vec<ColumnMetadata>,
/// None stands for not allocated yet.
table_route: Option<PhysicalTableRouteValue>,
/// None stands for not allocated yet.

View File

@@ -12,9 +12,11 @@
// See the License for the specific language governing permissions and
// limitations under the License.
use std::collections::HashSet;
use std::collections::{HashMap, HashSet};
use api::v1::SemanticType;
use common_telemetry::debug;
use common_telemetry::tracing::warn;
use store_api::metadata::ColumnMetadata;
use table::metadata::RawTableInfo;
@@ -23,6 +25,10 @@ pub(crate) fn build_new_physical_table_info(
mut raw_table_info: RawTableInfo,
physical_columns: &[ColumnMetadata],
) -> RawTableInfo {
debug!(
"building new physical table info for table: {}, table_id: {}",
raw_table_info.name, raw_table_info.ident.table_id
);
let existing_columns = raw_table_info
.meta
.schema
@@ -36,6 +42,8 @@ pub(crate) fn build_new_physical_table_info(
let time_index = &mut raw_table_info.meta.schema.timestamp_index;
let columns = &mut raw_table_info.meta.schema.column_schemas;
columns.clear();
let column_ids = &mut raw_table_info.meta.column_ids;
column_ids.clear();
for (idx, col) in physical_columns.iter().enumerate() {
match col.semantic_type {
@@ -50,6 +58,7 @@ pub(crate) fn build_new_physical_table_info(
}
columns.push(col.column_schema.clone());
column_ids.push(col.column_id);
}
if let Some(time_index) = *time_index {
@@ -58,3 +67,54 @@ pub(crate) fn build_new_physical_table_info(
raw_table_info
}
/// Updates the column IDs in the table info based on the provided column metadata.
///
/// This function validates that the column metadata matches the existing table schema
/// before updating the column ids. If the column metadata doesn't match the table schema,
/// the table info remains unchanged.
pub(crate) fn update_table_info_column_ids(
raw_table_info: &mut RawTableInfo,
column_metadatas: &[ColumnMetadata],
) {
let mut table_column_names = raw_table_info
.meta
.schema
.column_schemas
.iter()
.map(|c| c.name.as_str())
.collect::<Vec<_>>();
table_column_names.sort_unstable();
let mut column_names = column_metadatas
.iter()
.map(|c| c.column_schema.name.as_str())
.collect::<Vec<_>>();
column_names.sort_unstable();
if table_column_names != column_names {
warn!(
"Column metadata doesn't match the table schema for table {}, table_id: {}, column in table: {:?}, column in metadata: {:?}",
raw_table_info.name,
raw_table_info.ident.table_id,
table_column_names,
column_names,
);
return;
}
let name_to_id = column_metadatas
.iter()
.map(|c| (c.column_schema.name.clone(), c.column_id))
.collect::<HashMap<_, _>>();
let schema = &raw_table_info.meta.schema.column_schemas;
let mut column_ids = Vec::with_capacity(schema.len());
for column_schema in schema {
if let Some(id) = name_to_id.get(&column_schema.name) {
column_ids.push(*id);
}
}
raw_table_info.meta.column_ids = column_ids;
}

View File

@@ -122,6 +122,7 @@ impl TableMetadataAllocator {
);
let peers = self.peer_allocator.alloc(regions).await?;
debug!("Allocated peers {:?} for table {}", peers, table_id);
let region_routes = task
.partitions
.iter()

View File

@@ -24,7 +24,14 @@ use std::collections::HashMap;
use api::v1::meta::Partition;
use api::v1::{ColumnDataType, SemanticType};
use common_procedure::Status;
use store_api::metric_engine_consts::{LOGICAL_TABLE_METADATA_KEY, METRIC_ENGINE_NAME};
use datatypes::prelude::ConcreteDataType;
use datatypes::schema::ColumnSchema;
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::{
DATA_SCHEMA_TABLE_ID_COLUMN_NAME, DATA_SCHEMA_TSID_COLUMN_NAME, LOGICAL_TABLE_METADATA_KEY,
METRIC_ENGINE_NAME,
};
use store_api::storage::consts::ReservedColumnId;
use table::metadata::{RawTableInfo, TableId};
use crate::ddl::create_logical_tables::CreateLogicalTablesProcedure;
@@ -146,6 +153,7 @@ pub fn test_create_logical_table_task(name: &str) -> CreateTableTask {
}
}
/// Creates a physical table task with a single region.
pub fn test_create_physical_table_task(name: &str) -> CreateTableTask {
let create_table = TestCreateTableExprBuilder::default()
.column_defs([
@@ -182,3 +190,95 @@ pub fn test_create_physical_table_task(name: &str) -> CreateTableTask {
table_info,
}
}
/// Creates a column metadata list with tag fields.
pub fn test_column_metadatas(tag_fields: &[&str]) -> Vec<ColumnMetadata> {
let mut output = Vec::with_capacity(tag_fields.len() + 4);
output.extend([
ColumnMetadata {
column_schema: ColumnSchema::new(
"ts",
ConcreteDataType::timestamp_millisecond_datatype(),
false,
),
semantic_type: SemanticType::Timestamp,
column_id: 0,
},
ColumnMetadata {
column_schema: ColumnSchema::new("value", ConcreteDataType::float64_datatype(), false),
semantic_type: SemanticType::Field,
column_id: 1,
},
ColumnMetadata {
column_schema: ColumnSchema::new(
DATA_SCHEMA_TABLE_ID_COLUMN_NAME,
ConcreteDataType::timestamp_millisecond_datatype(),
false,
),
semantic_type: SemanticType::Tag,
column_id: ReservedColumnId::table_id(),
},
ColumnMetadata {
column_schema: ColumnSchema::new(
DATA_SCHEMA_TSID_COLUMN_NAME,
ConcreteDataType::float64_datatype(),
false,
),
semantic_type: SemanticType::Tag,
column_id: ReservedColumnId::tsid(),
},
]);
for (i, name) in tag_fields.iter().enumerate() {
output.push(ColumnMetadata {
column_schema: ColumnSchema::new(
name.to_string(),
ConcreteDataType::string_datatype(),
true,
),
semantic_type: SemanticType::Tag,
column_id: (i + 2) as u32,
});
}
output
}
/// Asserts the column names.
pub fn assert_column_name(table_info: &RawTableInfo, expected_column_names: &[&str]) {
assert_eq!(
table_info
.meta
.schema
.column_schemas
.iter()
.map(|c| c.name.to_string())
.collect::<Vec<_>>(),
expected_column_names
);
}
/// Asserts the column metadatas
pub fn assert_column_name_and_id(column_metadatas: &[ColumnMetadata], expected: &[(&str, u32)]) {
assert_eq!(expected.len(), column_metadatas.len());
for (name, id) in expected {
let column_metadata = column_metadatas
.iter()
.find(|c| c.column_id == *id)
.unwrap();
assert_eq!(column_metadata.column_schema.name, *name);
}
}
/// Gets the raw table info.
pub async fn get_raw_table_info(ddl_context: &DdlContext, table_id: TableId) -> RawTableInfo {
ddl_context
.table_metadata_manager
.table_info_manager()
.get(table_id)
.await
.unwrap()
.unwrap()
.into_inner()
.table_info
}

View File

@@ -132,6 +132,7 @@ pub fn build_raw_table_info_from_expr(expr: &CreateTableExpr) -> RawTableInfo {
options: TableOptions::try_from_iter(&expr.table_options).unwrap(),
created_on: DateTime::default(),
partition_key_indices: vec![],
column_ids: vec![],
},
table_type: TableType::Base,
}

View File

@@ -12,6 +12,8 @@
// See the License for the specific language governing permissions and
// limitations under the License.
use std::sync::Arc;
use api::region::RegionResponse;
use api::v1::region::RegionRequest;
use common_error::ext::{BoxedError, ErrorExt, StackError};
@@ -32,6 +34,7 @@ impl MockDatanodeHandler for () {
Ok(RegionResponse {
affected_rows: 0,
extensions: Default::default(),
metadata: Vec::new(),
})
}
@@ -44,10 +47,13 @@ impl MockDatanodeHandler for () {
}
}
type RegionRequestHandler =
Arc<dyn Fn(Peer, RegionRequest) -> Result<RegionResponse> + Send + Sync>;
#[derive(Clone)]
pub struct DatanodeWatcher {
sender: mpsc::Sender<(Peer, RegionRequest)>,
handler: Option<fn(Peer, RegionRequest) -> Result<RegionResponse>>,
handler: Option<RegionRequestHandler>,
}
impl DatanodeWatcher {
@@ -60,9 +66,9 @@ impl DatanodeWatcher {
pub fn with_handler(
mut self,
user_handler: fn(Peer, RegionRequest) -> Result<RegionResponse>,
user_handler: impl Fn(Peer, RegionRequest) -> Result<RegionResponse> + Send + Sync + 'static,
) -> Self {
self.handler = Some(user_handler);
self.handler = Some(Arc::new(user_handler));
self
}
}
@@ -75,7 +81,7 @@ impl MockDatanodeHandler for DatanodeWatcher {
.send((peer.clone(), request.clone()))
.await
.unwrap();
if let Some(handler) = self.handler {
if let Some(handler) = self.handler.as_ref() {
handler(peer.clone(), request)
} else {
Ok(RegionResponse::new(0))

View File

@@ -23,17 +23,20 @@ use api::v1::{ColumnDataType, SemanticType};
use common_catalog::consts::{DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME};
use common_procedure::{Procedure, ProcedureId, Status};
use common_procedure_test::MockContextProvider;
use store_api::metric_engine_consts::MANIFEST_INFO_EXTENSION_KEY;
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::{ALTER_PHYSICAL_EXTENSION_KEY, MANIFEST_INFO_EXTENSION_KEY};
use store_api::region_engine::RegionManifestInfo;
use store_api::storage::consts::ReservedColumnId;
use store_api::storage::RegionId;
use tokio::sync::mpsc;
use crate::ddl::alter_logical_tables::AlterLogicalTablesProcedure;
use crate::ddl::test_util::alter_table::TestAlterTableExprBuilder;
use crate::ddl::test_util::columns::TestColumnDefBuilder;
use crate::ddl::test_util::datanode_handler::{DatanodeWatcher, NaiveDatanodeHandler};
use crate::ddl::test_util::datanode_handler::DatanodeWatcher;
use crate::ddl::test_util::{
create_logical_table, create_physical_table, create_physical_table_metadata,
assert_column_name, create_logical_table, create_physical_table,
create_physical_table_metadata, get_raw_table_info, test_column_metadatas,
test_create_physical_table_task,
};
use crate::error::Error::{AlterLogicalTablesInvalidArguments, TableNotFound};
@@ -96,6 +99,52 @@ fn make_alter_logical_table_rename_task(
}
}
fn make_alters_request_handler(
column_metadatas: Vec<ColumnMetadata>,
) -> impl Fn(Peer, RegionRequest) -> Result<RegionResponse> {
move |_peer: Peer, request: RegionRequest| {
if let region_request::Body::Alters(_) = request.body.unwrap() {
let mut response = RegionResponse::new(0);
// Default region id for physical table.
let region_id = RegionId::new(1000, 1);
response.extensions.insert(
MANIFEST_INFO_EXTENSION_KEY.to_string(),
RegionManifestInfo::encode_list(&[(
region_id,
RegionManifestInfo::metric(1, 0, 2, 0),
)])
.unwrap(),
);
response.extensions.insert(
ALTER_PHYSICAL_EXTENSION_KEY.to_string(),
ColumnMetadata::encode_list(&column_metadatas).unwrap(),
);
return Ok(response);
}
Ok(RegionResponse::new(0))
}
}
fn assert_alters_request(
peer: Peer,
request: RegionRequest,
expected_peer_id: u64,
expected_region_ids: &[RegionId],
) {
assert_eq!(peer.id, expected_peer_id,);
let Some(region_request::Body::Alters(req)) = request.body else {
unreachable!();
};
for (i, region_id) in expected_region_ids.iter().enumerate() {
assert_eq!(
req.requests[i].region_id,
*region_id,
"actual region id: {}",
RegionId::from_u64(req.requests[i].region_id)
);
}
}
#[tokio::test]
async fn test_on_prepare_check_schema() {
let node_manager = Arc::new(MockDatanodeManager::new(()));
@@ -205,15 +254,20 @@ async fn test_on_prepare() {
#[tokio::test]
async fn test_on_update_metadata() {
let node_manager = Arc::new(MockDatanodeManager::new(NaiveDatanodeHandler));
common_telemetry::init_default_ut_logging();
let (tx, mut rx) = mpsc::channel(8);
let test_column_metadatas = test_column_metadatas(&["new_col", "mew_col"]);
let datanode_handler =
DatanodeWatcher::new(tx).with_handler(make_alters_request_handler(test_column_metadatas));
let node_manager = Arc::new(MockDatanodeManager::new(datanode_handler));
let ddl_context = new_ddl_context(node_manager);
// Creates physical table
let phy_id = create_physical_table(&ddl_context, "phy").await;
// Creates 3 logical tables
create_logical_table(ddl_context.clone(), phy_id, "table1").await;
create_logical_table(ddl_context.clone(), phy_id, "table2").await;
create_logical_table(ddl_context.clone(), phy_id, "table3").await;
let logical_table1_id = create_logical_table(ddl_context.clone(), phy_id, "table1").await;
let logical_table2_id = create_logical_table(ddl_context.clone(), phy_id, "table2").await;
let logical_table3_id = create_logical_table(ddl_context.clone(), phy_id, "table3").await;
create_logical_table(ddl_context.clone(), phy_id, "table4").await;
create_logical_table(ddl_context.clone(), phy_id, "table5").await;
@@ -223,7 +277,7 @@ async fn test_on_update_metadata() {
make_alter_logical_table_add_column_task(None, "table3", vec!["new_col".to_string()]),
];
let mut procedure = AlterLogicalTablesProcedure::new(tasks, phy_id, ddl_context);
let mut procedure = AlterLogicalTablesProcedure::new(tasks, phy_id, ddl_context.clone());
let mut status = procedure.on_prepare().await.unwrap();
assert_matches!(
status,
@@ -255,18 +309,52 @@ async fn test_on_update_metadata() {
clean_poisons: false
}
);
let (peer, request) = rx.try_recv().unwrap();
rx.try_recv().unwrap_err();
assert_alters_request(
peer,
request,
0,
&[
RegionId::new(logical_table1_id, 0),
RegionId::new(logical_table2_id, 0),
RegionId::new(logical_table3_id, 0),
],
);
let table_info = get_raw_table_info(&ddl_context, phy_id).await;
assert_column_name(
&table_info,
&["ts", "value", "__table_id", "__tsid", "new_col", "mew_col"],
);
assert_eq!(
table_info.meta.column_ids,
vec![
0,
1,
ReservedColumnId::table_id(),
ReservedColumnId::tsid(),
2,
3
]
);
}
#[tokio::test]
async fn test_on_part_duplicate_alter_request() {
let node_manager = Arc::new(MockDatanodeManager::new(NaiveDatanodeHandler));
let ddl_context = new_ddl_context(node_manager);
common_telemetry::init_default_ut_logging();
let (tx, mut rx) = mpsc::channel(8);
let column_metadatas = test_column_metadatas(&["col_0"]);
let handler =
DatanodeWatcher::new(tx).with_handler(make_alters_request_handler(column_metadatas));
let node_manager = Arc::new(MockDatanodeManager::new(handler));
let mut ddl_context = new_ddl_context(node_manager);
// Creates physical table
let phy_id = create_physical_table(&ddl_context, "phy").await;
// Creates 3 logical tables
create_logical_table(ddl_context.clone(), phy_id, "table1").await;
create_logical_table(ddl_context.clone(), phy_id, "table2").await;
let logical_table1_id = create_logical_table(ddl_context.clone(), phy_id, "table1").await;
let logical_table2_id = create_logical_table(ddl_context.clone(), phy_id, "table2").await;
let tasks = vec![
make_alter_logical_table_add_column_task(None, "table1", vec!["col_0".to_string()]),
@@ -305,6 +393,40 @@ async fn test_on_part_duplicate_alter_request() {
clean_poisons: false
}
);
let (peer, request) = rx.try_recv().unwrap();
rx.try_recv().unwrap_err();
assert_alters_request(
peer,
request,
0,
&[
RegionId::new(logical_table1_id, 0),
RegionId::new(logical_table2_id, 0),
],
);
let table_info = get_raw_table_info(&ddl_context, phy_id).await;
assert_column_name(
&table_info,
&["ts", "value", "__table_id", "__tsid", "col_0"],
);
assert_eq!(
table_info.meta.column_ids,
vec![
0,
1,
ReservedColumnId::table_id(),
ReservedColumnId::tsid(),
2
]
);
let (tx, mut rx) = mpsc::channel(8);
let column_metadatas = test_column_metadatas(&["col_0", "new_col_1", "new_col_2"]);
let handler =
DatanodeWatcher::new(tx).with_handler(make_alters_request_handler(column_metadatas));
let node_manager = Arc::new(MockDatanodeManager::new(handler));
ddl_context.node_manager = node_manager;
// re-alter
let tasks = vec![
@@ -357,6 +479,44 @@ async fn test_on_part_duplicate_alter_request() {
}
);
let (peer, request) = rx.try_recv().unwrap();
rx.try_recv().unwrap_err();
assert_alters_request(
peer,
request,
0,
&[
RegionId::new(logical_table1_id, 0),
RegionId::new(logical_table2_id, 0),
],
);
let table_info = get_raw_table_info(&ddl_context, phy_id).await;
assert_column_name(
&table_info,
&[
"ts",
"value",
"__table_id",
"__tsid",
"col_0",
"new_col_1",
"new_col_2",
],
);
assert_eq!(
table_info.meta.column_ids,
vec![
0,
1,
ReservedColumnId::table_id(),
ReservedColumnId::tsid(),
2,
3,
4,
]
);
let table_name_keys = vec![
TableNameKey::new(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, "table1"),
TableNameKey::new(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, "table2"),
@@ -422,27 +582,13 @@ async fn test_on_part_duplicate_alter_request() {
);
}
fn alters_request_handler(_peer: Peer, request: RegionRequest) -> Result<RegionResponse> {
if let region_request::Body::Alters(_) = request.body.unwrap() {
let mut response = RegionResponse::new(0);
// Default region id for physical table.
let region_id = RegionId::new(1000, 1);
response.extensions.insert(
MANIFEST_INFO_EXTENSION_KEY.to_string(),
RegionManifestInfo::encode_list(&[(region_id, RegionManifestInfo::metric(1, 0, 2, 0))])
.unwrap(),
);
return Ok(response);
}
Ok(RegionResponse::new(0))
}
#[tokio::test]
async fn test_on_submit_alter_region_request() {
common_telemetry::init_default_ut_logging();
let (tx, mut rx) = mpsc::channel(8);
let handler = DatanodeWatcher::new(tx).with_handler(alters_request_handler);
let column_metadatas = test_column_metadatas(&["new_col", "mew_col"]);
let handler =
DatanodeWatcher::new(tx).with_handler(make_alters_request_handler(column_metadatas));
let node_manager = Arc::new(MockDatanodeManager::new(handler));
let ddl_context = new_ddl_context(node_manager);

View File

@@ -30,7 +30,12 @@ use common_error::status_code::StatusCode;
use common_procedure::store::poison_store::PoisonStore;
use common_procedure::{ProcedureId, Status};
use common_procedure_test::MockContextProvider;
use store_api::metric_engine_consts::MANIFEST_INFO_EXTENSION_KEY;
use datatypes::prelude::ConcreteDataType;
use datatypes::schema::ColumnSchema;
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::{
MANIFEST_INFO_EXTENSION_KEY, TABLE_COLUMN_METADATA_EXTENSION_KEY,
};
use store_api::region_engine::RegionManifestInfo;
use store_api::storage::RegionId;
use table::requests::TTL_KEY;
@@ -43,6 +48,7 @@ use crate::ddl::test_util::datanode_handler::{
AllFailureDatanodeHandler, DatanodeWatcher, PartialSuccessDatanodeHandler,
RequestOutdatedErrorDatanodeHandler,
};
use crate::ddl::test_util::{assert_column_name, assert_column_name_and_id};
use crate::error::{Error, Result};
use crate::key::datanode_table::DatanodeTableKey;
use crate::key::table_name::TableNameKey;
@@ -179,6 +185,30 @@ fn alter_request_handler(_peer: Peer, request: RegionRequest) -> Result<RegionRe
RegionManifestInfo::encode_list(&[(region_id, RegionManifestInfo::mito(1, 1))])
.unwrap(),
);
response.extensions.insert(
TABLE_COLUMN_METADATA_EXTENSION_KEY.to_string(),
ColumnMetadata::encode_list(&[
ColumnMetadata {
column_schema: ColumnSchema::new(
"ts",
ConcreteDataType::timestamp_millisecond_datatype(),
false,
),
semantic_type: SemanticType::Timestamp,
column_id: 0,
},
ColumnMetadata {
column_schema: ColumnSchema::new(
"host",
ConcreteDataType::float64_datatype(),
false,
),
semantic_type: SemanticType::Tag,
column_id: 1,
},
])
.unwrap(),
);
return Ok(response);
}
@@ -187,6 +217,7 @@ fn alter_request_handler(_peer: Peer, request: RegionRequest) -> Result<RegionRe
#[tokio::test]
async fn test_on_submit_alter_request() {
common_telemetry::init_default_ut_logging();
let (tx, mut rx) = mpsc::channel(8);
let datanode_handler = DatanodeWatcher::new(tx).with_handler(alter_request_handler);
let node_manager = Arc::new(MockDatanodeManager::new(datanode_handler));
@@ -234,6 +265,8 @@ async fn test_on_submit_alter_request() {
assert_sync_request(peer, request, 4, RegionId::new(table_id, 2), 1);
let (peer, request) = results.remove(0);
assert_sync_request(peer, request, 5, RegionId::new(table_id, 1), 1);
let column_metadatas = procedure.data().column_metadatas();
assert_column_name_and_id(column_metadatas, &[("ts", 0), ("host", 1)]);
}
#[tokio::test]
@@ -378,6 +411,7 @@ async fn test_on_update_metadata_rename() {
#[tokio::test]
async fn test_on_update_metadata_add_columns() {
common_telemetry::init_default_ut_logging();
let node_manager = Arc::new(MockDatanodeManager::new(()));
let ddl_context = new_ddl_context(node_manager);
let table_name = "foo";
@@ -431,6 +465,34 @@ async fn test_on_update_metadata_add_columns() {
.submit_alter_region_requests(procedure_id, provider.as_ref())
.await
.unwrap();
// Returned column metadatas is empty.
assert!(procedure.data().column_metadatas().is_empty());
procedure.mut_data().set_column_metadatas(vec![
ColumnMetadata {
column_schema: ColumnSchema::new(
"ts",
ConcreteDataType::timestamp_millisecond_datatype(),
false,
),
semantic_type: SemanticType::Timestamp,
column_id: 0,
},
ColumnMetadata {
column_schema: ColumnSchema::new("host", ConcreteDataType::float64_datatype(), false),
semantic_type: SemanticType::Tag,
column_id: 1,
},
ColumnMetadata {
column_schema: ColumnSchema::new("cpu", ConcreteDataType::float64_datatype(), false),
semantic_type: SemanticType::Tag,
column_id: 2,
},
ColumnMetadata {
column_schema: ColumnSchema::new("my_tag3", ConcreteDataType::string_datatype(), true),
semantic_type: SemanticType::Tag,
column_id: 3,
},
]);
procedure.on_update_metadata().await.unwrap();
let table_info = ddl_context
@@ -447,6 +509,8 @@ async fn test_on_update_metadata_add_columns() {
table_info.meta.schema.column_schemas.len() as u32,
table_info.meta.next_column_id
);
assert_column_name(&table_info, &["ts", "host", "cpu", "my_tag3"]);
assert_eq!(table_info.meta.column_ids, vec![0, 1, 2, 3]);
}
#[tokio::test]

View File

@@ -141,3 +141,41 @@ async fn test_create_flow() {
let err = procedure.on_prepare().await.unwrap_err();
assert_matches!(err, error::Error::FlowAlreadyExists { .. });
}
#[tokio::test]
async fn test_create_flow_same_source_and_sink_table() {
let table_id = 1024;
let table_name = TableName::new(DEFAULT_CATALOG_NAME, DEFAULT_SCHEMA_NAME, "same_table");
// Use the same table for both source and sink
let source_table_names = vec![table_name.clone()];
let sink_table_name = table_name.clone();
let node_manager = Arc::new(MockFlownodeManager::new(NaiveFlownodeHandler));
let ddl_context = new_ddl_context(node_manager);
// Create the table first so it exists
let task = test_create_table_task("same_table", table_id);
ddl_context
.table_metadata_manager
.create_table_metadata(
task.table_info.clone(),
TableRouteValue::physical(vec![]),
HashMap::new(),
)
.await
.unwrap();
// Try to create a flow with same source and sink table - should fail
let task = test_create_flow_task("my_flow", source_table_names, sink_table_name, false);
let query_ctx = QueryContext::arc().into();
let mut procedure = CreateFlowProcedure::new(task, query_ctx, ddl_context);
let err = procedure.on_prepare().await.unwrap_err();
assert_matches!(err, error::Error::Unsupported { .. });
// Verify the error message contains information about the same table
if let error::Error::Unsupported { operation, .. } = &err {
assert!(operation.contains("source and sink table being the same"));
assert!(operation.contains("same_table"));
}
}

View File

@@ -23,15 +23,18 @@ use common_error::ext::ErrorExt;
use common_error::status_code::StatusCode;
use common_procedure::{Context as ProcedureContext, Procedure, ProcedureId, Status};
use common_procedure_test::MockContextProvider;
use store_api::metric_engine_consts::MANIFEST_INFO_EXTENSION_KEY;
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::{ALTER_PHYSICAL_EXTENSION_KEY, MANIFEST_INFO_EXTENSION_KEY};
use store_api::region_engine::RegionManifestInfo;
use store_api::storage::consts::ReservedColumnId;
use store_api::storage::RegionId;
use tokio::sync::mpsc;
use crate::ddl::create_logical_tables::CreateLogicalTablesProcedure;
use crate::ddl::test_util::datanode_handler::{DatanodeWatcher, NaiveDatanodeHandler};
use crate::ddl::test_util::{
create_physical_table_metadata, test_create_logical_table_task, test_create_physical_table_task,
assert_column_name, create_physical_table_metadata, get_raw_table_info, test_column_metadatas,
test_create_logical_table_task, test_create_physical_table_task,
};
use crate::ddl::TableMetadata;
use crate::error::{Error, Result};
@@ -39,6 +42,54 @@ use crate::key::table_route::{PhysicalTableRouteValue, TableRouteValue};
use crate::rpc::router::{Region, RegionRoute};
use crate::test_util::{new_ddl_context, MockDatanodeManager};
fn make_creates_request_handler(
column_metadatas: Vec<ColumnMetadata>,
) -> impl Fn(Peer, RegionRequest) -> Result<RegionResponse> {
move |_peer, request| {
let _ = _peer;
if let region_request::Body::Creates(_) = request.body.unwrap() {
let mut response = RegionResponse::new(0);
// Default region id for physical table.
let region_id = RegionId::new(1024, 1);
response.extensions.insert(
MANIFEST_INFO_EXTENSION_KEY.to_string(),
RegionManifestInfo::encode_list(&[(
region_id,
RegionManifestInfo::metric(1, 0, 2, 0),
)])
.unwrap(),
);
response.extensions.insert(
ALTER_PHYSICAL_EXTENSION_KEY.to_string(),
ColumnMetadata::encode_list(&column_metadatas).unwrap(),
);
return Ok(response);
}
Ok(RegionResponse::new(0))
}
}
fn assert_creates_request(
peer: Peer,
request: RegionRequest,
expected_peer_id: u64,
expected_region_ids: &[RegionId],
) {
assert_eq!(peer.id, expected_peer_id,);
let Some(region_request::Body::Creates(req)) = request.body else {
unreachable!();
};
for (i, region_id) in expected_region_ids.iter().enumerate() {
assert_eq!(
req.requests[i].region_id,
*region_id,
"actual region id: {}",
RegionId::from_u64(req.requests[i].region_id)
);
}
}
#[tokio::test]
async fn test_on_prepare_physical_table_not_found() {
let node_manager = Arc::new(MockDatanodeManager::new(()));
@@ -227,7 +278,12 @@ async fn test_on_prepare_part_logical_tables_exist() {
#[tokio::test]
async fn test_on_create_metadata() {
let node_manager = Arc::new(MockDatanodeManager::new(NaiveDatanodeHandler));
common_telemetry::init_default_ut_logging();
let (tx, mut rx) = mpsc::channel(8);
let column_metadatas = test_column_metadatas(&["host", "cpu"]);
let datanode_handler =
DatanodeWatcher::new(tx).with_handler(make_creates_request_handler(column_metadatas));
let node_manager = Arc::new(MockDatanodeManager::new(datanode_handler));
let ddl_context = new_ddl_context(node_manager);
// Prepares physical table metadata.
let mut create_physical_table_task = test_create_physical_table_task("phy_table");
@@ -255,7 +311,7 @@ async fn test_on_create_metadata() {
let mut procedure = CreateLogicalTablesProcedure::new(
vec![task, yet_another_task],
physical_table_id,
ddl_context,
ddl_context.clone(),
);
let status = procedure.on_prepare().await.unwrap();
assert_matches!(
@@ -274,11 +330,42 @@ async fn test_on_create_metadata() {
let status = procedure.execute(&ctx).await.unwrap();
let table_ids = status.downcast_output_ref::<Vec<u32>>().unwrap();
assert_eq!(*table_ids, vec![1025, 1026]);
let (peer, request) = rx.try_recv().unwrap();
rx.try_recv().unwrap_err();
assert_creates_request(
peer,
request,
0,
&[RegionId::new(1025, 0), RegionId::new(1026, 0)],
);
let table_info = get_raw_table_info(&ddl_context, table_id).await;
assert_column_name(
&table_info,
&["ts", "value", "__table_id", "__tsid", "host", "cpu"],
);
assert_eq!(
table_info.meta.column_ids,
vec![
0,
1,
ReservedColumnId::table_id(),
ReservedColumnId::tsid(),
2,
3
]
);
}
#[tokio::test]
async fn test_on_create_metadata_part_logical_tables_exist() {
let node_manager = Arc::new(MockDatanodeManager::new(NaiveDatanodeHandler));
common_telemetry::init_default_ut_logging();
let (tx, mut rx) = mpsc::channel(8);
let column_metadatas = test_column_metadatas(&["host", "cpu"]);
let datanode_handler =
DatanodeWatcher::new(tx).with_handler(make_creates_request_handler(column_metadatas));
let node_manager = Arc::new(MockDatanodeManager::new(datanode_handler));
let ddl_context = new_ddl_context(node_manager);
// Prepares physical table metadata.
let mut create_physical_table_task = test_create_physical_table_task("phy_table");
@@ -317,7 +404,7 @@ async fn test_on_create_metadata_part_logical_tables_exist() {
let mut procedure = CreateLogicalTablesProcedure::new(
vec![task, non_exist_task],
physical_table_id,
ddl_context,
ddl_context.clone(),
);
let status = procedure.on_prepare().await.unwrap();
assert_matches!(
@@ -336,6 +423,27 @@ async fn test_on_create_metadata_part_logical_tables_exist() {
let status = procedure.execute(&ctx).await.unwrap();
let table_ids = status.downcast_output_ref::<Vec<u32>>().unwrap();
assert_eq!(*table_ids, vec![8192, 1025]);
let (peer, request) = rx.try_recv().unwrap();
rx.try_recv().unwrap_err();
assert_creates_request(peer, request, 0, &[RegionId::new(1025, 0)]);
let table_info = get_raw_table_info(&ddl_context, table_id).await;
assert_column_name(
&table_info,
&["ts", "value", "__table_id", "__tsid", "host", "cpu"],
);
assert_eq!(
table_info.meta.column_ids,
vec![
0,
1,
ReservedColumnId::table_id(),
ReservedColumnId::tsid(),
2,
3
]
);
}
#[tokio::test]
@@ -399,27 +507,13 @@ async fn test_on_create_metadata_err() {
assert!(!error.is_retry_later());
}
fn creates_request_handler(_peer: Peer, request: RegionRequest) -> Result<RegionResponse> {
if let region_request::Body::Creates(_) = request.body.unwrap() {
let mut response = RegionResponse::new(0);
// Default region id for physical table.
let region_id = RegionId::new(1024, 1);
response.extensions.insert(
MANIFEST_INFO_EXTENSION_KEY.to_string(),
RegionManifestInfo::encode_list(&[(region_id, RegionManifestInfo::metric(1, 0, 2, 0))])
.unwrap(),
);
return Ok(response);
}
Ok(RegionResponse::new(0))
}
#[tokio::test]
async fn test_on_submit_create_request() {
common_telemetry::init_default_ut_logging();
let (tx, mut rx) = mpsc::channel(8);
let handler = DatanodeWatcher::new(tx).with_handler(creates_request_handler);
let column_metadatas = test_column_metadatas(&["host", "cpu"]);
let handler =
DatanodeWatcher::new(tx).with_handler(make_creates_request_handler(column_metadatas));
let node_manager = Arc::new(MockDatanodeManager::new(handler));
let ddl_context = new_ddl_context(node_manager);
let mut create_physical_table_task = test_create_physical_table_task("phy_table");

View File

@@ -16,7 +16,9 @@ use std::assert_matches::assert_matches;
use std::collections::HashMap;
use std::sync::Arc;
use api::v1::meta::Partition;
use api::region::RegionResponse;
use api::v1::meta::{Partition, Peer};
use api::v1::region::{region_request, RegionRequest};
use api::v1::{ColumnDataType, SemanticType};
use common_error::ext::ErrorExt;
use common_error::status_code::StatusCode;
@@ -24,7 +26,12 @@ use common_procedure::{Context as ProcedureContext, Procedure, ProcedureId, Stat
use common_procedure_test::{
execute_procedure_until, execute_procedure_until_done, MockContextProvider,
};
use datatypes::prelude::ConcreteDataType;
use datatypes::schema::ColumnSchema;
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::TABLE_COLUMN_METADATA_EXTENSION_KEY;
use store_api::storage::RegionId;
use tokio::sync::mpsc;
use crate::ddl::create_table::{CreateTableProcedure, CreateTableState};
use crate::ddl::test_util::columns::TestColumnDefBuilder;
@@ -32,14 +39,73 @@ use crate::ddl::test_util::create_table::{
build_raw_table_info_from_expr, TestCreateTableExprBuilder,
};
use crate::ddl::test_util::datanode_handler::{
NaiveDatanodeHandler, RetryErrorDatanodeHandler, UnexpectedErrorDatanodeHandler,
DatanodeWatcher, NaiveDatanodeHandler, RetryErrorDatanodeHandler,
UnexpectedErrorDatanodeHandler,
};
use crate::error::Error;
use crate::ddl::test_util::{assert_column_name, get_raw_table_info};
use crate::error::{Error, Result};
use crate::key::table_route::TableRouteValue;
use crate::kv_backend::memory::MemoryKvBackend;
use crate::rpc::ddl::CreateTableTask;
use crate::test_util::{new_ddl_context, new_ddl_context_with_kv_backend, MockDatanodeManager};
fn create_request_handler(_peer: Peer, request: RegionRequest) -> Result<RegionResponse> {
let _ = _peer;
if let region_request::Body::Create(_) = request.body.unwrap() {
let mut response = RegionResponse::new(0);
response.extensions.insert(
TABLE_COLUMN_METADATA_EXTENSION_KEY.to_string(),
ColumnMetadata::encode_list(&[
ColumnMetadata {
column_schema: ColumnSchema::new(
"ts",
ConcreteDataType::timestamp_millisecond_datatype(),
false,
),
semantic_type: SemanticType::Timestamp,
column_id: 0,
},
ColumnMetadata {
column_schema: ColumnSchema::new(
"host",
ConcreteDataType::float64_datatype(),
false,
),
semantic_type: SemanticType::Tag,
column_id: 1,
},
ColumnMetadata {
column_schema: ColumnSchema::new(
"cpu",
ConcreteDataType::float64_datatype(),
false,
),
semantic_type: SemanticType::Tag,
column_id: 2,
},
])
.unwrap(),
);
return Ok(response);
}
Ok(RegionResponse::new(0))
}
fn assert_create_request(
peer: Peer,
request: RegionRequest,
expected_peer_id: u64,
expected_region_id: RegionId,
) {
assert_eq!(peer.id, expected_peer_id);
let Some(region_request::Body::Create(req)) = request.body else {
unreachable!();
};
assert_eq!(req.region_id, expected_region_id);
}
pub(crate) fn test_create_table_task(name: &str) -> CreateTableTask {
let create_table = TestCreateTableExprBuilder::default()
.column_defs([
@@ -230,11 +296,13 @@ async fn test_on_create_metadata_error() {
#[tokio::test]
async fn test_on_create_metadata() {
common_telemetry::init_default_ut_logging();
let node_manager = Arc::new(MockDatanodeManager::new(NaiveDatanodeHandler));
let (tx, mut rx) = mpsc::channel(8);
let datanode_handler = DatanodeWatcher::new(tx).with_handler(create_request_handler);
let node_manager = Arc::new(MockDatanodeManager::new(datanode_handler));
let ddl_context = new_ddl_context(node_manager);
let task = test_create_table_task("foo");
assert!(!task.create_table.create_if_not_exists);
let mut procedure = CreateTableProcedure::new(task, ddl_context);
let mut procedure = CreateTableProcedure::new(task, ddl_context.clone());
procedure.on_prepare().await.unwrap();
let ctx = ProcedureContext {
procedure_id: ProcedureId::random(),
@@ -243,8 +311,16 @@ async fn test_on_create_metadata() {
procedure.execute(&ctx).await.unwrap();
// Triggers procedure to create table metadata
let status = procedure.execute(&ctx).await.unwrap();
let table_id = status.downcast_output_ref::<u32>().unwrap();
assert_eq!(*table_id, 1024);
let table_id = *status.downcast_output_ref::<u32>().unwrap();
assert_eq!(table_id, 1024);
let (peer, request) = rx.try_recv().unwrap();
rx.try_recv().unwrap_err();
assert_create_request(peer, request, 0, RegionId::new(table_id, 0));
let table_info = get_raw_table_info(&ddl_context, table_id).await;
assert_column_name(&table_info, &["ts", "host", "cpu"]);
assert_eq!(table_info.meta.column_ids, vec![0, 1, 2]);
}
#[tokio::test]

View File

@@ -29,6 +29,7 @@ use common_telemetry::{error, info, warn};
use common_wal::options::WalOptions;
use futures::future::join_all;
use snafu::{ensure, OptionExt, ResultExt};
use store_api::metadata::ColumnMetadata;
use store_api::metric_engine_consts::{LOGICAL_TABLE_METADATA_KEY, MANIFEST_INFO_EXTENSION_KEY};
use store_api::region_engine::RegionManifestInfo;
use store_api::storage::{RegionId, RegionNumber};
@@ -37,8 +38,8 @@ use table::table_reference::TableReference;
use crate::ddl::{DdlContext, DetectingRegion};
use crate::error::{
self, Error, OperateDatanodeSnafu, ParseWalOptionsSnafu, Result, TableNotFoundSnafu,
UnsupportedSnafu,
self, DecodeJsonSnafu, Error, MetadataCorruptionSnafu, OperateDatanodeSnafu,
ParseWalOptionsSnafu, Result, TableNotFoundSnafu, UnsupportedSnafu,
};
use crate::key::datanode_table::DatanodeTableValue;
use crate::key::table_name::TableNameKey;
@@ -314,11 +315,23 @@ pub fn parse_manifest_infos_from_extensions(
Ok(data_manifest_version)
}
/// Parses column metadatas from extensions.
pub fn parse_column_metadatas(
extensions: &HashMap<String, Vec<u8>>,
key: &str,
) -> Result<Vec<ColumnMetadata>> {
let value = extensions.get(key).context(error::UnexpectedSnafu {
err_msg: format!("column metadata extension not found: {}", key),
})?;
let column_metadatas = ColumnMetadata::decode_list(value).context(error::SerdeJsonSnafu {})?;
Ok(column_metadatas)
}
/// Sync follower regions on datanodes.
pub async fn sync_follower_regions(
context: &DdlContext,
table_id: TableId,
results: Vec<RegionResponse>,
results: &[RegionResponse],
region_routes: &[RegionRoute],
engine: &str,
) -> Result<()> {
@@ -331,7 +344,7 @@ pub async fn sync_follower_regions(
}
let results = results
.into_iter()
.iter()
.map(|response| parse_manifest_infos_from_extensions(&response.extensions))
.collect::<Result<Vec<_>>>()?
.into_iter()
@@ -418,6 +431,38 @@ pub async fn sync_follower_regions(
Ok(())
}
/// Extracts column metadatas from extensions.
pub fn extract_column_metadatas(
results: &mut [RegionResponse],
key: &str,
) -> Result<Option<Vec<ColumnMetadata>>> {
let schemas = results
.iter_mut()
.map(|r| r.extensions.remove(key))
.collect::<Vec<_>>();
if schemas.is_empty() {
return Ok(None);
}
// Verify all the physical schemas are the same
// Safety: previous check ensures this vec is not empty
let first = schemas.first().unwrap();
ensure!(
schemas.iter().all(|x| x == first),
MetadataCorruptionSnafu {
err_msg: "The table column metadata schemas from datanodes are not the same."
}
);
if let Some(first) = first {
let column_metadatas = ColumnMetadata::decode_list(first).context(DecodeJsonSnafu)?;
Ok(Some(column_metadatas))
} else {
Ok(None)
}
}
#[cfg(test)]
mod tests {
use super::*;

View File

@@ -50,7 +50,11 @@ use crate::key::{DeserializedValueWithBytes, TableMetadataManagerRef};
#[cfg(feature = "enterprise")]
use crate::rpc::ddl::trigger::CreateTriggerTask;
#[cfg(feature = "enterprise")]
use crate::rpc::ddl::trigger::DropTriggerTask;
#[cfg(feature = "enterprise")]
use crate::rpc::ddl::DdlTask::CreateTrigger;
#[cfg(feature = "enterprise")]
use crate::rpc::ddl::DdlTask::DropTrigger;
use crate::rpc::ddl::DdlTask::{
AlterDatabase, AlterLogicalTables, AlterTable, CreateDatabase, CreateFlow, CreateLogicalTables,
CreateTable, CreateView, DropDatabase, DropFlow, DropLogicalTables, DropTable, DropView,
@@ -91,6 +95,14 @@ pub trait TriggerDdlManager: Send + Sync {
query_context: QueryContext,
) -> Result<SubmitDdlTaskResponse>;
async fn drop_trigger(
&self,
drop_trigger_task: DropTriggerTask,
procedure_manager: ProcedureManagerRef,
ddl_context: DdlContext,
query_context: QueryContext,
) -> Result<SubmitDdlTaskResponse>;
fn as_any(&self) -> &dyn std::any::Any;
}
@@ -125,13 +137,12 @@ impl DdlManager {
ddl_context: DdlContext,
procedure_manager: ProcedureManagerRef,
register_loaders: bool,
#[cfg(feature = "enterprise")] trigger_ddl_manager: Option<TriggerDdlManagerRef>,
) -> Result<Self> {
let manager = Self {
ddl_context,
procedure_manager,
#[cfg(feature = "enterprise")]
trigger_ddl_manager,
trigger_ddl_manager: None,
};
if register_loaders {
manager.register_loaders()?;
@@ -139,6 +150,15 @@ impl DdlManager {
Ok(manager)
}
#[cfg(feature = "enterprise")]
pub fn with_trigger_ddl_manager(
mut self,
trigger_ddl_manager: Option<TriggerDdlManagerRef>,
) -> Self {
self.trigger_ddl_manager = trigger_ddl_manager;
self
}
/// Returns the [TableMetadataManagerRef].
pub fn table_metadata_manager(&self) -> &TableMetadataManagerRef {
&self.ddl_context.table_metadata_manager
@@ -640,6 +660,28 @@ async fn handle_drop_flow_task(
})
}
#[cfg(feature = "enterprise")]
async fn handle_drop_trigger_task(
ddl_manager: &DdlManager,
drop_trigger_task: DropTriggerTask,
query_context: QueryContext,
) -> Result<SubmitDdlTaskResponse> {
let Some(m) = ddl_manager.trigger_ddl_manager.as_ref() else {
return UnsupportedSnafu {
operation: "drop trigger",
}
.fail();
};
m.drop_trigger(
drop_trigger_task,
ddl_manager.procedure_manager.clone(),
ddl_manager.ddl_context.clone(),
query_context,
)
.await
}
async fn handle_drop_view_task(
ddl_manager: &DdlManager,
drop_view_task: DropViewTask,
@@ -827,6 +869,11 @@ impl ProcedureExecutor for DdlManager {
handle_create_flow_task(self, create_flow_task, request.query_context.into())
.await
}
DropFlow(drop_flow_task) => handle_drop_flow_task(self, drop_flow_task).await,
CreateView(create_view_task) => {
handle_create_view_task(self, create_view_task).await
}
DropView(drop_view_task) => handle_drop_view_task(self, drop_view_task).await,
#[cfg(feature = "enterprise")]
CreateTrigger(create_trigger_task) => {
handle_create_trigger_task(
@@ -836,11 +883,11 @@ impl ProcedureExecutor for DdlManager {
)
.await
}
DropFlow(drop_flow_task) => handle_drop_flow_task(self, drop_flow_task).await,
CreateView(create_view_task) => {
handle_create_view_task(self, create_view_task).await
#[cfg(feature = "enterprise")]
DropTrigger(drop_trigger_task) => {
handle_drop_trigger_task(self, drop_trigger_task, request.query_context.into())
.await
}
DropView(drop_view_task) => handle_drop_view_task(self, drop_view_task).await,
}
}
.trace(span)
@@ -948,6 +995,7 @@ mod tests {
Default::default(),
state_store,
poison_manager,
None,
));
let _ = DdlManager::try_new(
@@ -964,8 +1012,6 @@ mod tests {
},
procedure_manager.clone(),
true,
#[cfg(feature = "enterprise")]
None,
);
let expected_loaders = vec![

View File

@@ -100,8 +100,8 @@
pub mod catalog_name;
pub mod datanode_table;
pub mod flow;
pub mod maintenance;
pub mod node_address;
pub mod runtime_switch;
mod schema_metadata_manager;
pub mod schema_name;
pub mod table_info;
@@ -164,7 +164,9 @@ use crate::state_store::PoisonValue;
use crate::DatanodeId;
pub const NAME_PATTERN: &str = r"[a-zA-Z_:-][a-zA-Z0-9_:\-\.@#]*";
pub const MAINTENANCE_KEY: &str = "__maintenance";
pub const LEGACY_MAINTENANCE_KEY: &str = "__maintenance";
pub const MAINTENANCE_KEY: &str = "__switches/maintenance";
pub const PAUSE_PROCEDURE_KEY: &str = "__switches/pause_procedure";
pub const DATANODE_TABLE_KEY_PREFIX: &str = "__dn_table";
pub const TABLE_INFO_KEY_PREFIX: &str = "__table_info";

View File

@@ -1,86 +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::sync::Arc;
use crate::error::Result;
use crate::key::MAINTENANCE_KEY;
use crate::kv_backend::KvBackendRef;
use crate::rpc::store::PutRequest;
pub type MaintenanceModeManagerRef = Arc<MaintenanceModeManager>;
/// The maintenance mode manager.
///
/// Used to enable or disable maintenance mode.
#[derive(Clone)]
pub struct MaintenanceModeManager {
kv_backend: KvBackendRef,
}
impl MaintenanceModeManager {
pub fn new(kv_backend: KvBackendRef) -> Self {
Self { kv_backend }
}
/// Enables maintenance mode.
pub async fn set_maintenance_mode(&self) -> Result<()> {
let req = PutRequest {
key: Vec::from(MAINTENANCE_KEY),
value: vec![],
prev_kv: false,
};
self.kv_backend.put(req).await?;
Ok(())
}
/// Unsets maintenance mode.
pub async fn unset_maintenance_mode(&self) -> Result<()> {
self.kv_backend
.delete(MAINTENANCE_KEY.as_bytes(), false)
.await?;
Ok(())
}
/// Returns true if maintenance mode is enabled.
pub async fn maintenance_mode(&self) -> Result<bool> {
self.kv_backend.exists(MAINTENANCE_KEY.as_bytes()).await
}
}
#[cfg(test)]
mod tests {
use std::sync::Arc;
use crate::key::maintenance::MaintenanceModeManager;
use crate::kv_backend::memory::MemoryKvBackend;
#[tokio::test]
async fn test_maintenance_mode_manager() {
let maintenance_mode_manager = Arc::new(MaintenanceModeManager::new(Arc::new(
MemoryKvBackend::new(),
)));
assert!(!maintenance_mode_manager.maintenance_mode().await.unwrap());
maintenance_mode_manager
.set_maintenance_mode()
.await
.unwrap();
assert!(maintenance_mode_manager.maintenance_mode().await.unwrap());
maintenance_mode_manager
.unset_maintenance_mode()
.await
.unwrap();
assert!(!maintenance_mode_manager.maintenance_mode().await.unwrap());
}
}

View File

@@ -0,0 +1,224 @@
// 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 std::time::Duration;
use common_error::ext::BoxedError;
use common_procedure::local::PauseAware;
use moka::future::Cache;
use snafu::ResultExt;
use crate::error::{GetCacheSnafu, Result};
use crate::key::{LEGACY_MAINTENANCE_KEY, MAINTENANCE_KEY, PAUSE_PROCEDURE_KEY};
use crate::kv_backend::KvBackendRef;
use crate::rpc::store::{BatchDeleteRequest, PutRequest};
pub type RuntimeSwitchManagerRef = Arc<RuntimeSwitchManager>;
/// The runtime switch manager.
///
/// Used to enable or disable runtime switches.
#[derive(Clone)]
pub struct RuntimeSwitchManager {
kv_backend: KvBackendRef,
cache: Cache<Vec<u8>, Option<Vec<u8>>>,
}
#[async_trait::async_trait]
impl PauseAware for RuntimeSwitchManager {
async fn is_paused(&self) -> std::result::Result<bool, BoxedError> {
self.is_procedure_paused().await.map_err(BoxedError::new)
}
}
const CACHE_TTL: Duration = Duration::from_secs(10);
const MAX_CAPACITY: u64 = 32;
impl RuntimeSwitchManager {
pub fn new(kv_backend: KvBackendRef) -> Self {
let cache = Cache::builder()
.time_to_live(CACHE_TTL)
.max_capacity(MAX_CAPACITY)
.build();
Self { kv_backend, cache }
}
async fn put_key(&self, key: &str) -> Result<()> {
let req = PutRequest {
key: Vec::from(key),
value: vec![],
prev_kv: false,
};
self.kv_backend.put(req).await?;
self.cache.invalidate(key.as_bytes()).await;
Ok(())
}
async fn delete_keys(&self, keys: &[&str]) -> Result<()> {
let req = BatchDeleteRequest::new()
.with_keys(keys.iter().map(|x| x.as_bytes().to_vec()).collect());
self.kv_backend.batch_delete(req).await?;
for key in keys {
self.cache.invalidate(key.as_bytes()).await;
}
Ok(())
}
/// Returns true if the key exists.
async fn exists(&self, key: &str) -> Result<bool> {
let key = key.as_bytes().to_vec();
let kv_backend = self.kv_backend.clone();
let value = self
.cache
.try_get_with(key.clone(), async move {
kv_backend.get(&key).await.map(|v| v.map(|v| v.value))
})
.await
.context(GetCacheSnafu)?;
Ok(value.is_some())
}
/// Enables maintenance mode.
pub async fn set_maintenance_mode(&self) -> Result<()> {
self.put_key(MAINTENANCE_KEY).await
}
/// Unsets maintenance mode.
pub async fn unset_maintenance_mode(&self) -> Result<()> {
self.delete_keys(&[MAINTENANCE_KEY, LEGACY_MAINTENANCE_KEY])
.await
}
/// Returns true if maintenance mode is enabled.
pub async fn maintenance_mode(&self) -> Result<bool> {
let exists = self.exists(MAINTENANCE_KEY).await?;
if exists {
return Ok(true);
}
let exists = self.exists(LEGACY_MAINTENANCE_KEY).await?;
if exists {
return Ok(true);
}
Ok(false)
}
// Pauses handling of incoming procedure requests.
pub async fn pasue_procedure(&self) -> Result<()> {
self.put_key(PAUSE_PROCEDURE_KEY).await
}
/// Resumes processing of incoming procedure requests.
pub async fn resume_procedure(&self) -> Result<()> {
self.delete_keys(&[PAUSE_PROCEDURE_KEY]).await
}
/// Returns true if the system is currently pausing incoming procedure requests.
pub async fn is_procedure_paused(&self) -> Result<bool> {
self.exists(PAUSE_PROCEDURE_KEY).await
}
}
#[cfg(test)]
mod tests {
use std::sync::Arc;
use crate::key::runtime_switch::RuntimeSwitchManager;
use crate::key::{LEGACY_MAINTENANCE_KEY, MAINTENANCE_KEY};
use crate::kv_backend::memory::MemoryKvBackend;
use crate::kv_backend::KvBackend;
use crate::rpc::store::PutRequest;
#[tokio::test]
async fn test_runtime_switch_manager_basic() {
let runtime_switch_manager =
Arc::new(RuntimeSwitchManager::new(Arc::new(MemoryKvBackend::new())));
runtime_switch_manager
.put_key(MAINTENANCE_KEY)
.await
.unwrap();
let v = runtime_switch_manager
.cache
.get(MAINTENANCE_KEY.as_bytes())
.await;
assert!(v.is_none());
runtime_switch_manager
.exists(MAINTENANCE_KEY)
.await
.unwrap();
let v = runtime_switch_manager
.cache
.get(MAINTENANCE_KEY.as_bytes())
.await;
assert!(v.is_some());
runtime_switch_manager
.delete_keys(&[MAINTENANCE_KEY])
.await
.unwrap();
let v = runtime_switch_manager
.cache
.get(MAINTENANCE_KEY.as_bytes())
.await;
assert!(v.is_none());
}
#[tokio::test]
async fn test_runtime_switch_manager() {
let runtime_switch_manager =
Arc::new(RuntimeSwitchManager::new(Arc::new(MemoryKvBackend::new())));
assert!(!runtime_switch_manager.maintenance_mode().await.unwrap());
runtime_switch_manager.set_maintenance_mode().await.unwrap();
assert!(runtime_switch_manager.maintenance_mode().await.unwrap());
runtime_switch_manager
.unset_maintenance_mode()
.await
.unwrap();
assert!(!runtime_switch_manager.maintenance_mode().await.unwrap());
}
#[tokio::test]
async fn test_runtime_switch_manager_with_legacy_key() {
let kv_backend = Arc::new(MemoryKvBackend::new());
kv_backend
.put(PutRequest {
key: Vec::from(LEGACY_MAINTENANCE_KEY),
value: vec![],
prev_kv: false,
})
.await
.unwrap();
let runtime_switch_manager = Arc::new(RuntimeSwitchManager::new(kv_backend));
assert!(runtime_switch_manager.maintenance_mode().await.unwrap());
runtime_switch_manager
.unset_maintenance_mode()
.await
.unwrap();
assert!(!runtime_switch_manager.maintenance_mode().await.unwrap());
runtime_switch_manager.set_maintenance_mode().await.unwrap();
assert!(runtime_switch_manager.maintenance_mode().await.unwrap());
}
#[tokio::test]
async fn test_pasue_procedure() {
let runtime_switch_manager =
Arc::new(RuntimeSwitchManager::new(Arc::new(MemoryKvBackend::new())));
runtime_switch_manager.pasue_procedure().await.unwrap();
assert!(runtime_switch_manager.is_procedure_paused().await.unwrap());
runtime_switch_manager.resume_procedure().await.unwrap();
assert!(!runtime_switch_manager.is_procedure_paused().await.unwrap());
}
}

View File

@@ -334,6 +334,7 @@ mod tests {
options: Default::default(),
region_numbers: vec![1],
partition_key_indices: vec![],
column_ids: vec![],
};
RawTableInfo {

View File

@@ -48,6 +48,11 @@ impl TableRouteKey {
pub fn new(table_id: TableId) -> Self {
Self { table_id }
}
/// Returns the range prefix of the table route key.
pub fn range_prefix() -> Vec<u8> {
format!("{}/", TABLE_ROUTE_PREFIX).into_bytes()
}
}
#[derive(Debug, PartialEq, Serialize, Deserialize, Clone)]

View File

@@ -14,13 +14,14 @@
use std::collections::HashMap;
use common_telemetry::debug;
use snafu::ensure;
use crate::error::{self, Result};
use crate::key::txn_helper::TxnOpGetResponseSet;
use crate::kv_backend::txn::{Compare, CompareOp, Txn, TxnOp};
use crate::kv_backend::KvBackendRef;
use crate::rpc::store::BatchGetRequest;
use crate::rpc::store::{BatchDeleteRequest, BatchGetRequest};
/// [TombstoneManager] provides the ability to:
/// - logically delete values
@@ -28,6 +29,9 @@ use crate::rpc::store::BatchGetRequest;
pub struct TombstoneManager {
kv_backend: KvBackendRef,
tombstone_prefix: String,
// Only used for testing.
#[cfg(test)]
max_txn_ops: Option<usize>,
}
const TOMBSTONE_PREFIX: &str = "__tombstone/";
@@ -35,10 +39,7 @@ const TOMBSTONE_PREFIX: &str = "__tombstone/";
impl TombstoneManager {
/// Returns [TombstoneManager].
pub fn new(kv_backend: KvBackendRef) -> Self {
Self {
kv_backend,
tombstone_prefix: TOMBSTONE_PREFIX.to_string(),
}
Self::new_with_prefix(kv_backend, TOMBSTONE_PREFIX)
}
/// Returns [TombstoneManager] with a custom tombstone prefix.
@@ -46,6 +47,8 @@ impl TombstoneManager {
Self {
kv_backend,
tombstone_prefix: prefix.to_string(),
#[cfg(test)]
max_txn_ops: None,
}
}
@@ -53,6 +56,11 @@ impl TombstoneManager {
[self.tombstone_prefix.as_bytes(), key].concat()
}
#[cfg(test)]
pub fn set_max_txn_ops(&mut self, max_txn_ops: usize) {
self.max_txn_ops = Some(max_txn_ops);
}
/// Moves value to `dest_key`.
///
/// Puts `value` to `dest_key` if the value of `src_key` equals `value`.
@@ -83,7 +91,11 @@ impl TombstoneManager {
ensure!(
keys.len() == dest_keys.len(),
error::UnexpectedSnafu {
err_msg: "The length of keys does not match the length of dest_keys."
err_msg: format!(
"The length of keys({}) does not match the length of dest_keys({}).",
keys.len(),
dest_keys.len()
),
}
);
// The key -> dest key mapping.
@@ -136,19 +148,45 @@ impl TombstoneManager {
.fail()
}
fn max_txn_ops(&self) -> usize {
#[cfg(test)]
if let Some(max_txn_ops) = self.max_txn_ops {
return max_txn_ops;
}
self.kv_backend.max_txn_ops()
}
/// Moves values to `dest_key`.
///
/// Returns the number of keys that were moved.
async fn move_values(&self, keys: Vec<Vec<u8>>, dest_keys: Vec<Vec<u8>>) -> Result<usize> {
let chunk_size = self.kv_backend.max_txn_ops() / 2;
if keys.len() > chunk_size {
let keys_chunks = keys.chunks(chunk_size).collect::<Vec<_>>();
let dest_keys_chunks = keys.chunks(chunk_size).collect::<Vec<_>>();
for (keys, dest_keys) in keys_chunks.into_iter().zip(dest_keys_chunks) {
self.move_values_inner(keys, dest_keys).await?;
ensure!(
keys.len() == dest_keys.len(),
error::UnexpectedSnafu {
err_msg: format!(
"The length of keys({}) does not match the length of dest_keys({}).",
keys.len(),
dest_keys.len()
),
}
Ok(keys.len())
);
if keys.is_empty() {
return Ok(0);
}
let chunk_size = self.max_txn_ops() / 2;
if keys.len() > chunk_size {
debug!(
"Moving values with multiple chunks, keys len: {}, chunk_size: {}",
keys.len(),
chunk_size
);
let mut moved_keys = 0;
let keys_chunks = keys.chunks(chunk_size).collect::<Vec<_>>();
let dest_keys_chunks = dest_keys.chunks(chunk_size).collect::<Vec<_>>();
for (keys, dest_keys) in keys_chunks.into_iter().zip(dest_keys_chunks) {
moved_keys += self.move_values_inner(keys, dest_keys).await?;
}
Ok(moved_keys)
} else {
self.move_values_inner(&keys, &dest_keys).await
}
@@ -196,15 +234,18 @@ impl TombstoneManager {
///
/// Returns the number of keys that were deleted.
pub async fn delete(&self, keys: Vec<Vec<u8>>) -> Result<usize> {
let operations = keys
let keys = keys
.iter()
.map(|key| TxnOp::Delete(self.to_tombstone(key)))
.map(|key| self.to_tombstone(key))
.collect::<Vec<_>>();
let txn = Txn::new().and_then(operations);
// Always success.
let _ = self.kv_backend.txn(txn).await?;
Ok(keys.len())
let num_keys = keys.len();
let _ = self
.kv_backend
.batch_delete(BatchDeleteRequest::new().with_keys(keys))
.await?;
Ok(num_keys)
}
}
@@ -392,16 +433,73 @@ mod tests {
.into_iter()
.map(|kv| (kv.key, kv.dest_key))
.unzip();
tombstone_manager
let moved_keys = tombstone_manager
.move_values(keys.clone(), dest_keys.clone())
.await
.unwrap();
assert_eq!(kvs.len(), moved_keys);
check_moved_values(kv_backend.clone(), &move_values).await;
// Moves again
tombstone_manager
let moved_keys = tombstone_manager
.move_values(keys.clone(), dest_keys.clone())
.await
.unwrap();
assert_eq!(0, moved_keys);
check_moved_values(kv_backend.clone(), &move_values).await;
}
#[tokio::test]
async fn test_move_values_with_max_txn_ops() {
common_telemetry::init_default_ut_logging();
let kv_backend = Arc::new(MemoryKvBackend::default());
let mut tombstone_manager = TombstoneManager::new(kv_backend.clone());
tombstone_manager.set_max_txn_ops(4);
let kvs = HashMap::from([
(b"bar".to_vec(), b"baz".to_vec()),
(b"foo".to_vec(), b"hi".to_vec()),
(b"baz".to_vec(), b"hello".to_vec()),
(b"qux".to_vec(), b"world".to_vec()),
(b"quux".to_vec(), b"world".to_vec()),
(b"quuux".to_vec(), b"world".to_vec()),
(b"quuuux".to_vec(), b"world".to_vec()),
(b"quuuuux".to_vec(), b"world".to_vec()),
(b"quuuuuux".to_vec(), b"world".to_vec()),
]);
for (key, value) in &kvs {
kv_backend
.put(
PutRequest::new()
.with_key(key.clone())
.with_value(value.clone()),
)
.await
.unwrap();
}
let move_values = kvs
.iter()
.map(|(key, value)| MoveValue {
key: key.clone(),
dest_key: tombstone_manager.to_tombstone(key),
value: value.clone(),
})
.collect::<Vec<_>>();
let (keys, dest_keys): (Vec<_>, Vec<_>) = move_values
.clone()
.into_iter()
.map(|kv| (kv.key, kv.dest_key))
.unzip();
let moved_keys = tombstone_manager
.move_values(keys.clone(), dest_keys.clone())
.await
.unwrap();
assert_eq!(kvs.len(), moved_keys);
check_moved_values(kv_backend.clone(), &move_values).await;
// Moves again
let moved_keys = tombstone_manager
.move_values(keys.clone(), dest_keys.clone())
.await
.unwrap();
assert_eq!(0, moved_keys);
check_moved_values(kv_backend.clone(), &move_values).await;
}
@@ -439,17 +537,19 @@ mod tests {
.unzip();
keys.push(b"non-exists".to_vec());
dest_keys.push(b"hi/non-exists".to_vec());
tombstone_manager
let moved_keys = tombstone_manager
.move_values(keys.clone(), dest_keys.clone())
.await
.unwrap();
check_moved_values(kv_backend.clone(), &move_values).await;
assert_eq!(3, moved_keys);
// Moves again
tombstone_manager
let moved_keys = tombstone_manager
.move_values(keys.clone(), dest_keys.clone())
.await
.unwrap();
check_moved_values(kv_backend.clone(), &move_values).await;
assert_eq!(0, moved_keys);
}
#[tokio::test]
@@ -490,10 +590,11 @@ mod tests {
.into_iter()
.map(|kv| (kv.key, kv.dest_key))
.unzip();
tombstone_manager
let moved_keys = tombstone_manager
.move_values(keys, dest_keys)
.await
.unwrap();
assert_eq!(kvs.len(), moved_keys);
}
#[tokio::test]
@@ -571,4 +672,24 @@ mod tests {
.unwrap();
check_moved_values(kv_backend.clone(), &move_values).await;
}
#[tokio::test]
async fn test_move_values_with_different_lengths() {
let kv_backend = Arc::new(MemoryKvBackend::default());
let tombstone_manager = TombstoneManager::new(kv_backend.clone());
let keys = vec![b"bar".to_vec(), b"foo".to_vec()];
let dest_keys = vec![b"bar".to_vec(), b"foo".to_vec(), b"baz".to_vec()];
let err = tombstone_manager
.move_values(keys, dest_keys)
.await
.unwrap_err();
assert!(err
.to_string()
.contains("The length of keys(2) does not match the length of dest_keys(3)."),);
let moved_keys = tombstone_manager.move_values(vec![], vec![]).await.unwrap();
assert_eq!(0, moved_keys);
}
}

View File

@@ -69,6 +69,8 @@ pub enum DdlTask {
AlterDatabase(AlterDatabaseTask),
CreateFlow(CreateFlowTask),
DropFlow(DropFlowTask),
#[cfg(feature = "enterprise")]
DropTrigger(trigger::DropTriggerTask),
CreateView(CreateViewTask),
DropView(DropViewTask),
#[cfg(feature = "enterprise")]
@@ -259,6 +261,18 @@ impl TryFrom<Task> for DdlTask {
.fail()
}
}
Task::DropTriggerTask(drop_trigger) => {
#[cfg(feature = "enterprise")]
return Ok(DdlTask::DropTrigger(drop_trigger.try_into()?));
#[cfg(not(feature = "enterprise"))]
{
let _ = drop_trigger;
crate::error::UnsupportedSnafu {
operation: "drop trigger",
}
.fail()
}
}
}
}
}
@@ -311,6 +325,8 @@ impl TryFrom<SubmitDdlTaskRequest> for PbDdlTaskRequest {
DdlTask::DropView(task) => Task::DropViewTask(task.into()),
#[cfg(feature = "enterprise")]
DdlTask::CreateTrigger(task) => Task::CreateTriggerTask(task.into()),
#[cfg(feature = "enterprise")]
DdlTask::DropTrigger(task) => Task::DropTriggerTask(task.into()),
};
Ok(Self {
@@ -1358,6 +1374,7 @@ mod tests {
options: Default::default(),
created_on: Default::default(),
partition_key_indices: Default::default(),
column_ids: Default::default(),
};
// construct RawTableInfo

View File

@@ -1,10 +1,13 @@
use std::collections::HashMap;
use std::time::Duration;
use api::v1::meta::CreateTriggerTask as PbCreateTriggerTask;
use api::v1::meta::{
CreateTriggerTask as PbCreateTriggerTask, DropTriggerTask as PbDropTriggerTask,
};
use api::v1::notify_channel::ChannelType as PbChannelType;
use api::v1::{
CreateTriggerExpr, NotifyChannel as PbNotifyChannel, WebhookOptions as PbWebhookOptions,
CreateTriggerExpr as PbCreateTriggerExpr, DropTriggerExpr as PbDropTriggerExpr,
NotifyChannel as PbNotifyChannel, WebhookOptions as PbWebhookOptions,
};
use serde::{Deserialize, Serialize};
use snafu::OptionExt;
@@ -56,7 +59,7 @@ impl From<CreateTriggerTask> for PbCreateTriggerTask {
.map(PbNotifyChannel::from)
.collect();
let expr = CreateTriggerExpr {
let expr = PbCreateTriggerExpr {
catalog_name: task.catalog_name,
trigger_name: task.trigger_name,
create_if_not_exists: task.if_not_exists,
@@ -139,17 +142,86 @@ impl TryFrom<PbNotifyChannel> for NotifyChannel {
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct DropTriggerTask {
pub catalog_name: String,
pub trigger_name: String,
pub drop_if_exists: bool,
}
impl From<DropTriggerTask> for PbDropTriggerTask {
fn from(task: DropTriggerTask) -> Self {
let expr = PbDropTriggerExpr {
catalog_name: task.catalog_name,
trigger_name: task.trigger_name,
drop_if_exists: task.drop_if_exists,
};
PbDropTriggerTask {
drop_trigger: Some(expr),
}
}
}
impl TryFrom<PbDropTriggerTask> for DropTriggerTask {
type Error = error::Error;
fn try_from(task: PbDropTriggerTask) -> Result<Self> {
let expr = task.drop_trigger.context(error::InvalidProtoMsgSnafu {
err_msg: "expected drop_trigger",
})?;
Ok(DropTriggerTask {
catalog_name: expr.catalog_name,
trigger_name: expr.trigger_name,
drop_if_exists: expr.drop_if_exists,
})
}
}
impl DdlTask {
/// Creates a [`DdlTask`] to create a trigger.
pub fn new_create_trigger(expr: CreateTriggerTask) -> Self {
DdlTask::CreateTrigger(expr)
}
/// Creates a [`DdlTask`] to drop a trigger.
pub fn new_drop_trigger(expr: DropTriggerTask) -> Self {
DdlTask::DropTrigger(expr)
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_convert_drop_trigger_task() {
let original = DropTriggerTask {
catalog_name: "test_catalog".to_string(),
trigger_name: "test_trigger".to_string(),
drop_if_exists: true,
};
let pb_task: PbDropTriggerTask = original.clone().into();
let expr = pb_task.drop_trigger.as_ref().unwrap();
assert_eq!(expr.catalog_name, "test_catalog");
assert_eq!(expr.trigger_name, "test_trigger");
assert!(expr.drop_if_exists);
let round_tripped = DropTriggerTask::try_from(pb_task).unwrap();
assert_eq!(original.catalog_name, round_tripped.catalog_name);
assert_eq!(original.trigger_name, round_tripped.trigger_name);
assert_eq!(original.drop_if_exists, round_tripped.drop_if_exists);
// Test invalid case where drop_trigger is None
let invalid_task = PbDropTriggerTask { drop_trigger: None };
let result = DropTriggerTask::try_from(invalid_task);
assert!(result.is_err());
}
#[test]
fn test_convert_create_trigger_task() {
let original = CreateTriggerTask {

View File

@@ -28,6 +28,19 @@ use crate::PoisonKey;
#[snafu(visibility(pub))]
#[stack_trace_debug]
pub enum Error {
#[snafu(display("Failed to check procedure manager status"))]
CheckStatus {
source: BoxedError,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Manager is pasued"))]
ManagerPasued {
#[snafu(implicit)]
location: Location,
},
#[snafu(display(
"Failed to execute procedure due to external error, clean poisons: {}",
clean_poisons
@@ -246,7 +259,8 @@ impl ErrorExt for Error {
| Error::ListState { source, .. }
| Error::PutPoison { source, .. }
| Error::DeletePoison { source, .. }
| Error::GetPoison { source, .. } => source.status_code(),
| Error::GetPoison { source, .. }
| Error::CheckStatus { source, .. } => source.status_code(),
Error::ToJson { .. }
| Error::DeleteState { .. }
@@ -259,7 +273,8 @@ impl ErrorExt for Error {
Error::RetryTimesExceeded { .. }
| Error::RollbackTimesExceeded { .. }
| Error::ManagerNotStart { .. } => StatusCode::IllegalState,
| Error::ManagerNotStart { .. }
| Error::ManagerPasued { .. } => StatusCode::IllegalState,
Error::RollbackNotSupported { .. } => StatusCode::Unsupported,
Error::LoaderConflict { .. } | Error::DuplicateProcedure { .. } => {

View File

@@ -22,6 +22,7 @@ use std::time::{Duration, Instant};
use async_trait::async_trait;
use backon::ExponentialBuilder;
use common_error::ext::BoxedError;
use common_runtime::{RepeatedTask, TaskFunction};
use common_telemetry::tracing_context::{FutureExt, TracingContext};
use common_telemetry::{error, info, tracing};
@@ -30,9 +31,10 @@ use tokio::sync::watch::{self, Receiver, Sender};
use tokio::sync::{Mutex as TokioMutex, Notify};
use crate::error::{
self, DuplicateProcedureSnafu, Error, LoaderConflictSnafu, ManagerNotStartSnafu,
PoisonKeyNotDefinedSnafu, ProcedureNotFoundSnafu, Result, StartRemoveOutdatedMetaTaskSnafu,
StopRemoveOutdatedMetaTaskSnafu, TooManyRunningProceduresSnafu,
self, CheckStatusSnafu, DuplicateProcedureSnafu, Error, LoaderConflictSnafu,
ManagerNotStartSnafu, ManagerPasuedSnafu, PoisonKeyNotDefinedSnafu, ProcedureNotFoundSnafu,
Result, StartRemoveOutdatedMetaTaskSnafu, StopRemoveOutdatedMetaTaskSnafu,
TooManyRunningProceduresSnafu,
};
use crate::local::runner::Runner;
use crate::procedure::{BoxedProcedureLoader, InitProcedureState, PoisonKeys, ProcedureInfo};
@@ -522,6 +524,14 @@ impl Default for ManagerConfig {
}
}
type PauseAwareRef = Arc<dyn PauseAware>;
#[async_trait]
pub trait PauseAware: Send + Sync {
/// Returns true if the procedure manager is paused.
async fn is_paused(&self) -> std::result::Result<bool, BoxedError>;
}
/// A [ProcedureManager] that maintains procedure states locally.
pub struct LocalManager {
manager_ctx: Arc<ManagerContext>,
@@ -531,6 +541,7 @@ pub struct LocalManager {
/// GC task.
remove_outdated_meta_task: TokioMutex<Option<RepeatedTask<Error>>>,
config: ManagerConfig,
pause_aware: Option<PauseAwareRef>,
}
impl LocalManager {
@@ -539,6 +550,7 @@ impl LocalManager {
config: ManagerConfig,
state_store: StateStoreRef,
poison_store: PoisonStoreRef,
pause_aware: Option<PauseAwareRef>,
) -> LocalManager {
let manager_ctx = Arc::new(ManagerContext::new(poison_store));
@@ -549,6 +561,7 @@ impl LocalManager {
retry_delay: config.retry_delay,
remove_outdated_meta_task: TokioMutex::new(None),
config,
pause_aware,
}
}
@@ -719,6 +732,17 @@ impl LocalManager {
let loaders = self.manager_ctx.loaders.lock().unwrap();
loaders.contains_key(name)
}
async fn check_status(&self) -> Result<()> {
if let Some(pause_aware) = self.pause_aware.as_ref() {
ensure!(
!pause_aware.is_paused().await.context(CheckStatusSnafu)?,
ManagerPasuedSnafu
);
}
Ok(())
}
}
#[async_trait]
@@ -774,6 +798,7 @@ impl ProcedureManager for LocalManager {
!self.manager_ctx.contains_procedure(procedure_id),
DuplicateProcedureSnafu { procedure_id }
);
self.check_status().await?;
self.submit_root(
procedure.id,
@@ -979,7 +1004,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(test_util::new_object_store(&dir)));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.manager_ctx.start();
manager
@@ -1004,7 +1029,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(object_store.clone()));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.manager_ctx.start();
manager
@@ -1058,7 +1083,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(test_util::new_object_store(&dir)));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.manager_ctx.start();
let procedure_id = ProcedureId::random();
@@ -1110,7 +1135,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(test_util::new_object_store(&dir)));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.manager_ctx.start();
#[derive(Debug)]
@@ -1191,7 +1216,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(test_util::new_object_store(&dir)));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
let mut procedure = ProcedureToLoad::new("submit");
procedure.lock_key = LockKey::single_exclusive("test.submit");
@@ -1219,7 +1244,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(test_util::new_object_store(&dir)));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.start().await.unwrap();
manager.stop().await.unwrap();
@@ -1256,7 +1281,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(object_store.clone()));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.manager_ctx.set_running();
let mut procedure = ProcedureToLoad::new("submit");
@@ -1338,7 +1363,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(test_util::new_object_store(&dir)));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.manager_ctx.set_running();
manager
@@ -1463,7 +1488,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(object_store.clone()));
let poison_manager = Arc::new(InMemoryPoisonStore::new());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.manager_ctx.start();
let notify = Arc::new(Notify::new());

View File

@@ -83,7 +83,7 @@ mod tests {
};
let state_store = Arc::new(ObjectStateStore::new(test_util::new_object_store(&dir)));
let poison_manager = Arc::new(InMemoryPoisonStore::default());
let manager = LocalManager::new(config, state_store, poison_manager);
let manager = LocalManager::new(config, state_store, poison_manager, None);
manager.start().await.unwrap();
#[derive(Debug)]

View File

@@ -14,7 +14,7 @@
pub mod columnar_value;
pub mod error;
mod function;
pub mod function;
pub mod logical_plan;
pub mod prelude;
pub mod request;

View File

@@ -178,8 +178,6 @@ pub enum Error {
StreamTimeout {
#[snafu(implicit)]
location: Location,
#[snafu(source)]
error: tokio::time::error::Elapsed,
},
#[snafu(display("RecordBatch slice index overflow: {visit_index} > {size}"))]

22
src/common/sql/Cargo.toml Normal file
View File

@@ -0,0 +1,22 @@
[package]
name = "common-sql"
version.workspace = true
edition.workspace = true
license.workspace = true
[dependencies]
common-base.workspace = true
common-datasource.workspace = true
common-decimal.workspace = true
common-error.workspace = true
common-macro.workspace = true
common-time.workspace = true
datafusion-sql.workspace = true
datatypes.workspace = true
hex = "0.4"
jsonb.workspace = true
snafu.workspace = true
sqlparser.workspace = true
[lints]
workspace = true

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,182 @@
// 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_time::timezone::Timezone;
use datatypes::prelude::ConcreteDataType;
use datatypes::schema::constraint::{CURRENT_TIMESTAMP, CURRENT_TIMESTAMP_FN};
use datatypes::schema::ColumnDefaultConstraint;
pub use sqlparser::ast::{
visit_expressions_mut, visit_statements_mut, BinaryOperator, ColumnDef, ColumnOption,
ColumnOptionDef, DataType, Expr, Function, FunctionArg, FunctionArgExpr, FunctionArguments,
Ident, ObjectName, SqlOption, TableConstraint, TimezoneInfo, UnaryOperator, Value as SqlValue,
Visit, VisitMut, Visitor, VisitorMut,
};
use crate::convert::{sql_number_to_value, sql_value_to_value};
use crate::error::{Result, UnsupportedDefaultValueSnafu};
pub fn parse_column_default_constraint(
column_name: &str,
data_type: &ConcreteDataType,
opts: &[ColumnOptionDef],
timezone: Option<&Timezone>,
) -> Result<Option<ColumnDefaultConstraint>> {
if let Some(opt) = opts
.iter()
.find(|o| matches!(o.option, ColumnOption::Default(_)))
{
let default_constraint = match &opt.option {
ColumnOption::Default(Expr::Value(v)) => ColumnDefaultConstraint::Value(
sql_value_to_value(column_name, data_type, v, timezone, None, false)?,
),
ColumnOption::Default(Expr::Function(func)) => {
let mut func = format!("{func}").to_lowercase();
// normalize CURRENT_TIMESTAMP to CURRENT_TIMESTAMP()
if func == CURRENT_TIMESTAMP {
func = CURRENT_TIMESTAMP_FN.to_string();
}
// Always use lowercase for function expression
ColumnDefaultConstraint::Function(func.to_lowercase())
}
ColumnOption::Default(Expr::UnaryOp { op, expr }) => {
// Specialized process for handling numerical inputs to prevent
// overflow errors during the parsing of negative numbers,
// See https://github.com/GreptimeTeam/greptimedb/issues/4351
if let (UnaryOperator::Minus, Expr::Value(SqlValue::Number(n, _))) =
(op, expr.as_ref())
{
return Ok(Some(ColumnDefaultConstraint::Value(sql_number_to_value(
data_type,
&format!("-{n}"),
)?)));
}
if let Expr::Value(v) = &**expr {
let value =
sql_value_to_value(column_name, data_type, v, timezone, Some(*op), false)?;
ColumnDefaultConstraint::Value(value)
} else {
return UnsupportedDefaultValueSnafu {
column_name,
expr: *expr.clone(),
}
.fail();
}
}
ColumnOption::Default(others) => {
return UnsupportedDefaultValueSnafu {
column_name,
expr: others.clone(),
}
.fail();
}
_ => {
return UnsupportedDefaultValueSnafu {
column_name,
expr: Expr::Value(SqlValue::Null),
}
.fail();
}
};
Ok(Some(default_constraint))
} else {
Ok(None)
}
}
#[cfg(test)]
mod test {
use std::assert_matches::assert_matches;
use datatypes::prelude::{ConcreteDataType, Value};
use datatypes::types::BooleanType;
use super::*;
#[test]
pub fn test_parse_column_default_constraint() {
let bool_value = sqlparser::ast::Value::Boolean(true);
let opts = vec![
ColumnOptionDef {
name: None,
option: ColumnOption::Default(Expr::Value(bool_value)),
},
ColumnOptionDef {
name: None,
option: ColumnOption::NotNull,
},
];
let constraint = parse_column_default_constraint(
"coll",
&ConcreteDataType::Boolean(BooleanType),
&opts,
None,
)
.unwrap();
assert_matches!(
constraint,
Some(ColumnDefaultConstraint::Value(Value::Boolean(true)))
);
// Test negative number
let opts = vec![ColumnOptionDef {
name: None,
option: ColumnOption::Default(Expr::UnaryOp {
op: UnaryOperator::Minus,
expr: Box::new(Expr::Value(SqlValue::Number("32768".to_string(), false))),
}),
}];
let constraint = parse_column_default_constraint(
"coll",
&ConcreteDataType::int16_datatype(),
&opts,
None,
)
.unwrap();
assert_matches!(
constraint,
Some(ColumnDefaultConstraint::Value(Value::Int16(-32768)))
);
}
#[test]
fn test_incorrect_default_value_issue_3479() {
let opts = vec![ColumnOptionDef {
name: None,
option: ColumnOption::Default(Expr::Value(SqlValue::Number(
"0.047318541668048164".into(),
false,
))),
}];
let constraint = parse_column_default_constraint(
"coll",
&ConcreteDataType::float64_datatype(),
&opts,
None,
)
.unwrap()
.unwrap();
assert_eq!("0.047318541668048164", constraint.to_string());
let encoded: Vec<u8> = constraint.clone().try_into().unwrap();
let decoded = ColumnDefaultConstraint::try_from(encoded.as_ref()).unwrap();
assert_eq!(decoded, constraint);
}
}

158
src/common/sql/src/error.rs Normal file
View File

@@ -0,0 +1,158 @@
// 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 common_error::ext::ErrorExt;
use common_error::status_code::StatusCode;
use common_macro::stack_trace_debug;
use common_time::timestamp::TimeUnit;
use common_time::Timestamp;
use datafusion_sql::sqlparser::ast::UnaryOperator;
use datatypes::prelude::{ConcreteDataType, Value};
use snafu::{Location, Snafu};
pub use sqlparser::ast::{Expr, Value as SqlValue};
pub type Result<T> = std::result::Result<T, Error>;
/// SQL parser errors.
// Now the error in parser does not contain backtrace to avoid generating backtrace
// every time the parser parses an invalid SQL.
#[derive(Snafu)]
#[snafu(visibility(pub))]
#[stack_trace_debug]
pub enum Error {
#[snafu(display(
"Column {} expect type: {:?}, actual: {:?}",
column_name,
expect,
actual,
))]
ColumnTypeMismatch {
column_name: String,
expect: ConcreteDataType,
actual: ConcreteDataType,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Failed to parse value: {}", msg))]
ParseSqlValue {
msg: String,
#[snafu(implicit)]
location: Location,
},
#[snafu(display(
"Unsupported expr in default constraint: {:?} for column: {}",
expr,
column_name
))]
UnsupportedDefaultValue {
column_name: String,
expr: Expr,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Unable to convert sql value {} to datatype {:?}", value, datatype))]
ConvertSqlValue {
value: SqlValue,
datatype: ConcreteDataType,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Invalid sql value: {}", value))]
InvalidSqlValue {
value: String,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Unsupported unary operator {}", unary_op))]
UnsupportedUnaryOp {
unary_op: UnaryOperator,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Invalid unary operator {} for value {}", unary_op, value))]
InvalidUnaryOp {
unary_op: UnaryOperator,
value: Value,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Failed to cast SQL value {} to datatype {}", sql_value, datatype))]
InvalidCast {
sql_value: sqlparser::ast::Value,
datatype: ConcreteDataType,
#[snafu(implicit)]
location: Location,
source: datatypes::error::Error,
},
#[snafu(display("Unable to convert {} to datatype {:?}", value, datatype))]
ConvertStr {
value: String,
datatype: ConcreteDataType,
#[snafu(implicit)]
location: Location,
},
#[snafu(display(
"Converting timestamp {:?} to unit {:?} overflow",
timestamp,
target_unit
))]
TimestampOverflow {
timestamp: Timestamp,
target_unit: TimeUnit,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Datatype error: {}", source))]
Datatype {
source: datatypes::error::Error,
#[snafu(implicit)]
location: Location,
},
}
impl ErrorExt for Error {
fn status_code(&self) -> StatusCode {
use Error::*;
match self {
UnsupportedDefaultValue { .. } => StatusCode::Unsupported,
ParseSqlValue { .. } => StatusCode::InvalidSyntax,
ColumnTypeMismatch { .. }
| InvalidSqlValue { .. }
| UnsupportedUnaryOp { .. }
| InvalidUnaryOp { .. }
| InvalidCast { .. }
| ConvertStr { .. }
| TimestampOverflow { .. } => StatusCode::InvalidArguments,
Datatype { source, .. } => source.status_code(),
ConvertSqlValue { .. } => StatusCode::Unsupported,
}
}
fn as_any(&self) -> &dyn Any {
self
}
}

19
src/common/sql/src/lib.rs Normal file
View File

@@ -0,0 +1,19 @@
// 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)]
pub mod convert;
pub mod default_constraint;
pub mod error;

View File

@@ -22,7 +22,7 @@ once_cell.workspace = true
opentelemetry = { version = "0.21.0", default-features = false, features = [
"trace",
] }
opentelemetry-otlp = { version = "0.14.0", features = ["tokio"] }
opentelemetry-otlp = { version = "0.14.0", features = ["tokio", "http-proto", "reqwest-client"] }
opentelemetry-semantic-conventions = "0.13.0"
opentelemetry_sdk = { version = "0.21.0", features = ["rt-tokio"] }
parking_lot.workspace = true

View File

@@ -20,7 +20,7 @@ use std::time::Duration;
use once_cell::sync::{Lazy, OnceCell};
use opentelemetry::{global, KeyValue};
use opentelemetry_otlp::WithExportConfig;
use opentelemetry_otlp::{Protocol, SpanExporterBuilder, WithExportConfig};
use opentelemetry_sdk::propagation::TraceContextPropagator;
use opentelemetry_sdk::trace::Sampler;
use opentelemetry_semantic_conventions::resource;
@@ -36,7 +36,11 @@ use tracing_subscriber::{filter, EnvFilter, Registry};
use crate::tracing_sampler::{create_sampler, TracingSampleOptions};
pub const DEFAULT_OTLP_ENDPOINT: &str = "http://localhost:4317";
/// The default endpoint when use gRPC exporter protocol.
pub const DEFAULT_OTLP_GRPC_ENDPOINT: &str = "http://localhost:4317";
/// The default endpoint when use HTTP exporter protocol.
pub const DEFAULT_OTLP_HTTP_ENDPOINT: &str = "http://localhost:4318";
/// The default logs directory.
pub const DEFAULT_LOGGING_DIR: &str = "logs";
@@ -67,11 +71,25 @@ pub struct LoggingOptions {
/// Whether to enable tracing with OTLP. Default is false.
pub enable_otlp_tracing: bool,
/// The endpoint of OTLP. Default is "http://localhost:4317".
/// The endpoint of OTLP. Default is "http://localhost:4318".
pub otlp_endpoint: Option<String>,
/// The tracing sample ratio.
pub tracing_sample_ratio: Option<TracingSampleOptions>,
/// The protocol of OTLP export.
pub otlp_export_protocol: Option<OtlpExportProtocol>,
}
/// The protocol of OTLP export.
#[derive(Clone, Debug, Serialize, Deserialize, PartialEq)]
#[serde(rename_all = "snake_case")]
pub enum OtlpExportProtocol {
/// GRPC protocol.
Grpc,
/// HTTP protocol with binary protobuf.
Http,
}
/// The options of slow query.
@@ -147,6 +165,7 @@ impl Default for LoggingOptions {
append_stdout: true,
// Rotation hourly, 24 files per day, keeps info log files of 30 days
max_log_files: 720,
otlp_export_protocol: None,
}
}
}
@@ -388,22 +407,9 @@ pub fn init_global_logging(
KeyValue::new(resource::PROCESS_PID, std::process::id().to_string()),
]));
let exporter = opentelemetry_otlp::new_exporter().tonic().with_endpoint(
opts.otlp_endpoint
.as_ref()
.map(|e| {
if e.starts_with("http") {
e.to_string()
} else {
format!("http://{}", e)
}
})
.unwrap_or(DEFAULT_OTLP_ENDPOINT.to_string()),
);
let tracer = opentelemetry_otlp::new_pipeline()
.tracing()
.with_exporter(exporter)
.with_exporter(build_otlp_exporter(opts))
.with_trace_config(trace_config)
.install_batch(opentelemetry_sdk::runtime::Tokio)
.expect("otlp tracer install failed");
@@ -421,6 +427,42 @@ pub fn init_global_logging(
guards
}
fn build_otlp_exporter(opts: &LoggingOptions) -> SpanExporterBuilder {
let protocol = opts
.otlp_export_protocol
.clone()
.unwrap_or(OtlpExportProtocol::Http);
let endpoint = opts
.otlp_endpoint
.as_ref()
.map(|e| {
if e.starts_with("http") {
e.to_string()
} else {
format!("http://{}", e)
}
})
.unwrap_or_else(|| match protocol {
OtlpExportProtocol::Grpc => DEFAULT_OTLP_GRPC_ENDPOINT.to_string(),
OtlpExportProtocol::Http => DEFAULT_OTLP_HTTP_ENDPOINT.to_string(),
});
match protocol {
OtlpExportProtocol::Grpc => SpanExporterBuilder::Tonic(
opentelemetry_otlp::new_exporter()
.tonic()
.with_endpoint(endpoint),
),
OtlpExportProtocol::Http => SpanExporterBuilder::Http(
opentelemetry_otlp::new_exporter()
.http()
.with_endpoint(endpoint)
.with_protocol(Protocol::HttpBinary),
),
}
}
fn build_slow_query_logger<S>(
opts: &LoggingOptions,
slow_query_opts: Option<&SlowQueryOptions>,

View File

@@ -86,26 +86,33 @@ impl RegionAliveKeeper {
/// Add the countdown task for a specific region.
/// It will be ignored if the task exists.
pub async fn register_region(&self, region_id: RegionId) {
if self.find_handle(region_id).await.is_some() {
return;
}
let handle = Arc::new(CountdownTaskHandle::new(
self.region_server.clone(),
self.countdown_task_handler_ext.clone(),
region_id,
));
let mut handles = self.tasks.lock().await;
let _ = handles.insert(region_id, handle.clone());
let should_start = {
let mut handles = self.tasks.lock().await;
if self.started.load(Ordering::Relaxed) {
// Check if already exists, return early if so
if handles.contains_key(&region_id) {
return;
}
// Insert new handle
handles.insert(region_id, handle.clone());
// Return whether we should start (check state inside lock)
self.started.load(Ordering::Relaxed)
};
if should_start {
handle.start(self.heartbeat_interval_millis).await;
info!("Region alive countdown for region {region_id} is started!",);
info!("Region alive countdown for region {region_id} is started!");
} else {
info!(
"Region alive countdown for region {region_id} is registered but not started yet!",
"Region alive countdown for region {region_id} is registered but not started yet!"
);
}
}
@@ -475,7 +482,7 @@ mod test {
async fn region_alive_keeper() {
common_telemetry::init_default_ut_logging();
let mut region_server = mock_region_server();
let mut engine_env = TestEnv::with_prefix("region-alive-keeper");
let mut engine_env = TestEnv::with_prefix("region-alive-keeper").await;
let engine = engine_env.create_engine(MitoConfig::default()).await;
let engine = Arc::new(engine);
region_server.register_engine(engine.clone());

View File

@@ -14,10 +14,7 @@
//! Datanode configurations
use core::time::Duration;
use common_base::readable_size::ReadableSize;
use common_base::secrets::{ExposeSecret, SecretString};
use common_config::{Configurable, DEFAULT_DATA_HOME};
pub use common_procedure::options::ProcedureConfig;
use common_telemetry::logging::{LoggingOptions, TracingOptions};
@@ -27,6 +24,7 @@ use file_engine::config::EngineConfig as FileEngineConfig;
use meta_client::MetaClientOptions;
use metric_engine::config::EngineConfig as MetricEngineConfig;
use mito2::config::MitoConfig;
pub(crate) use object_store::config::ObjectStoreConfig;
use query::options::QueryOptions;
use serde::{Deserialize, Serialize};
use servers::export_metrics::ExportMetricsOption;
@@ -36,53 +34,6 @@ use servers::http::HttpOptions;
pub const DEFAULT_OBJECT_STORE_CACHE_SIZE: ReadableSize = ReadableSize::gb(5);
/// Object storage config
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
#[serde(tag = "type")]
pub enum ObjectStoreConfig {
File(FileConfig),
S3(S3Config),
Oss(OssConfig),
Azblob(AzblobConfig),
Gcs(GcsConfig),
}
impl ObjectStoreConfig {
/// Returns the object storage type name, such as `S3`, `Oss` etc.
pub fn provider_name(&self) -> &'static str {
match self {
Self::File(_) => "File",
Self::S3(_) => "S3",
Self::Oss(_) => "Oss",
Self::Azblob(_) => "Azblob",
Self::Gcs(_) => "Gcs",
}
}
/// Returns true when it's a remote object storage such as AWS s3 etc.
pub fn is_object_storage(&self) -> bool {
!matches!(self, Self::File(_))
}
/// Returns the object storage configuration name, return the provider name if it's empty.
pub fn config_name(&self) -> &str {
let name = match self {
// file storage doesn't support name
Self::File(_) => self.provider_name(),
Self::S3(s3) => &s3.name,
Self::Oss(oss) => &oss.name,
Self::Azblob(az) => &az.name,
Self::Gcs(gcs) => &gcs.name,
};
if name.trim().is_empty() {
return self.provider_name();
}
name
}
}
/// Storage engine config
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
#[serde(default)]
@@ -112,252 +63,6 @@ impl Default for StorageConfig {
}
}
#[derive(Debug, Clone, Serialize, Default, Deserialize, Eq, PartialEq)]
#[serde(default)]
pub struct FileConfig {}
#[derive(Debug, Clone, Serialize, Deserialize, Default, PartialEq)]
#[serde(default)]
pub struct ObjectStorageCacheConfig {
/// The local file cache directory
pub cache_path: Option<String>,
/// The cache capacity in bytes
pub cache_capacity: Option<ReadableSize>,
}
/// The http client options to the storage.
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
#[serde(default)]
pub struct HttpClientConfig {
/// The maximum idle connection per host allowed in the pool.
pub(crate) pool_max_idle_per_host: u32,
/// The timeout for only the connect phase of a http client.
#[serde(with = "humantime_serde")]
pub(crate) connect_timeout: Duration,
/// The total request timeout, applied from when the request starts connecting until the response body has finished.
/// Also considered a total deadline.
#[serde(with = "humantime_serde")]
pub(crate) timeout: Duration,
/// The timeout for idle sockets being kept-alive.
#[serde(with = "humantime_serde")]
pub(crate) pool_idle_timeout: Duration,
/// Skip SSL certificate validation (insecure)
pub skip_ssl_validation: bool,
}
impl Default for HttpClientConfig {
fn default() -> Self {
Self {
pool_max_idle_per_host: 1024,
connect_timeout: Duration::from_secs(30),
timeout: Duration::from_secs(30),
pool_idle_timeout: Duration::from_secs(90),
skip_ssl_validation: false,
}
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
#[serde(default)]
pub struct S3Config {
pub name: String,
pub bucket: String,
pub root: String,
#[serde(skip_serializing)]
pub access_key_id: SecretString,
#[serde(skip_serializing)]
pub secret_access_key: SecretString,
pub endpoint: Option<String>,
pub region: Option<String>,
/// Enable virtual host style so that opendal will send API requests in virtual host style instead of path style.
/// By default, opendal will send API to https://s3.us-east-1.amazonaws.com/bucket_name
/// Enabled, opendal will send API to https://bucket_name.s3.us-east-1.amazonaws.com
pub enable_virtual_host_style: bool,
#[serde(flatten)]
pub cache: ObjectStorageCacheConfig,
pub http_client: HttpClientConfig,
}
impl PartialEq for S3Config {
fn eq(&self, other: &Self) -> bool {
self.name == other.name
&& self.bucket == other.bucket
&& self.root == other.root
&& self.access_key_id.expose_secret() == other.access_key_id.expose_secret()
&& self.secret_access_key.expose_secret() == other.secret_access_key.expose_secret()
&& self.endpoint == other.endpoint
&& self.region == other.region
&& self.enable_virtual_host_style == other.enable_virtual_host_style
&& self.cache == other.cache
&& self.http_client == other.http_client
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
#[serde(default)]
pub struct OssConfig {
pub name: String,
pub bucket: String,
pub root: String,
#[serde(skip_serializing)]
pub access_key_id: SecretString,
#[serde(skip_serializing)]
pub access_key_secret: SecretString,
pub endpoint: String,
#[serde(flatten)]
pub cache: ObjectStorageCacheConfig,
pub http_client: HttpClientConfig,
}
impl PartialEq for OssConfig {
fn eq(&self, other: &Self) -> bool {
self.name == other.name
&& self.bucket == other.bucket
&& self.root == other.root
&& self.access_key_id.expose_secret() == other.access_key_id.expose_secret()
&& self.access_key_secret.expose_secret() == other.access_key_secret.expose_secret()
&& self.endpoint == other.endpoint
&& self.cache == other.cache
&& self.http_client == other.http_client
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
#[serde(default)]
pub struct AzblobConfig {
pub name: String,
pub container: String,
pub root: String,
#[serde(skip_serializing)]
pub account_name: SecretString,
#[serde(skip_serializing)]
pub account_key: SecretString,
pub endpoint: String,
pub sas_token: Option<String>,
#[serde(flatten)]
pub cache: ObjectStorageCacheConfig,
pub http_client: HttpClientConfig,
}
impl PartialEq for AzblobConfig {
fn eq(&self, other: &Self) -> bool {
self.name == other.name
&& self.container == other.container
&& self.root == other.root
&& self.account_name.expose_secret() == other.account_name.expose_secret()
&& self.account_key.expose_secret() == other.account_key.expose_secret()
&& self.endpoint == other.endpoint
&& self.sas_token == other.sas_token
&& self.cache == other.cache
&& self.http_client == other.http_client
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
#[serde(default)]
pub struct GcsConfig {
pub name: String,
pub root: String,
pub bucket: String,
pub scope: String,
#[serde(skip_serializing)]
pub credential_path: SecretString,
#[serde(skip_serializing)]
pub credential: SecretString,
pub endpoint: String,
#[serde(flatten)]
pub cache: ObjectStorageCacheConfig,
pub http_client: HttpClientConfig,
}
impl PartialEq for GcsConfig {
fn eq(&self, other: &Self) -> bool {
self.name == other.name
&& self.root == other.root
&& self.bucket == other.bucket
&& self.scope == other.scope
&& self.credential_path.expose_secret() == other.credential_path.expose_secret()
&& self.credential.expose_secret() == other.credential.expose_secret()
&& self.endpoint == other.endpoint
&& self.cache == other.cache
&& self.http_client == other.http_client
}
}
impl Default for S3Config {
fn default() -> Self {
Self {
name: String::default(),
bucket: String::default(),
root: String::default(),
access_key_id: SecretString::from(String::default()),
secret_access_key: SecretString::from(String::default()),
enable_virtual_host_style: false,
endpoint: Option::default(),
region: Option::default(),
cache: ObjectStorageCacheConfig::default(),
http_client: HttpClientConfig::default(),
}
}
}
impl Default for OssConfig {
fn default() -> Self {
Self {
name: String::default(),
bucket: String::default(),
root: String::default(),
access_key_id: SecretString::from(String::default()),
access_key_secret: SecretString::from(String::default()),
endpoint: String::default(),
cache: ObjectStorageCacheConfig::default(),
http_client: HttpClientConfig::default(),
}
}
}
impl Default for AzblobConfig {
fn default() -> Self {
Self {
name: String::default(),
container: String::default(),
root: String::default(),
account_name: SecretString::from(String::default()),
account_key: SecretString::from(String::default()),
endpoint: String::default(),
sas_token: Option::default(),
cache: ObjectStorageCacheConfig::default(),
http_client: HttpClientConfig::default(),
}
}
}
impl Default for GcsConfig {
fn default() -> Self {
Self {
name: String::default(),
root: String::default(),
bucket: String::default(),
scope: String::default(),
credential_path: SecretString::from(String::default()),
credential: SecretString::from(String::default()),
endpoint: String::default(),
cache: ObjectStorageCacheConfig::default(),
http_client: HttpClientConfig::default(),
}
}
}
impl Default for ObjectStoreConfig {
fn default() -> Self {
ObjectStoreConfig::File(FileConfig {})
}
}
#[derive(Clone, Debug, Serialize, Deserialize, PartialEq)]
#[serde(default)]
pub struct DatanodeOptions {
@@ -467,37 +172,6 @@ mod tests {
let _parsed: DatanodeOptions = toml::from_str(&toml_string).unwrap();
}
#[test]
fn test_config_name() {
let object_store_config = ObjectStoreConfig::default();
assert_eq!("File", object_store_config.config_name());
let s3_config = ObjectStoreConfig::S3(S3Config::default());
assert_eq!("S3", s3_config.config_name());
assert_eq!("S3", s3_config.provider_name());
let s3_config = ObjectStoreConfig::S3(S3Config {
name: "test".to_string(),
..Default::default()
});
assert_eq!("test", s3_config.config_name());
assert_eq!("S3", s3_config.provider_name());
}
#[test]
fn test_is_object_storage() {
let store = ObjectStoreConfig::default();
assert!(!store.is_object_storage());
let s3_config = ObjectStoreConfig::S3(S3Config::default());
assert!(s3_config.is_object_storage());
let oss_config = ObjectStoreConfig::Oss(OssConfig::default());
assert!(oss_config.is_object_storage());
let gcs_config = ObjectStoreConfig::Gcs(GcsConfig::default());
assert!(gcs_config.is_object_storage());
let azblob_config = ObjectStoreConfig::Azblob(AzblobConfig::default());
assert!(azblob_config.is_object_storage());
}
#[test]
fn test_secstr() {
let toml_str = r#"

View File

@@ -40,7 +40,7 @@ use log_store::raft_engine::log_store::RaftEngineLogStore;
use meta_client::MetaClientRef;
use metric_engine::engine::MetricEngine;
use mito2::config::MitoConfig;
use mito2::engine::MitoEngine;
use mito2::engine::{MitoEngine, MitoEngineBuilder};
use object_store::manager::{ObjectStoreManager, ObjectStoreManagerRef};
use object_store::util::normalize_dir;
use query::dummy_catalog::TableProviderFactoryRef;
@@ -162,6 +162,8 @@ pub struct DatanodeBuilder {
meta_client: Option<MetaClientRef>,
kv_backend: KvBackendRef,
cache_registry: Option<Arc<LayeredCacheRegistry>>,
#[cfg(feature = "enterprise")]
extension_range_provider_factory: Option<mito2::extension::BoxedExtensionRangeProviderFactory>,
}
impl DatanodeBuilder {
@@ -173,6 +175,8 @@ impl DatanodeBuilder {
meta_client: None,
kv_backend,
cache_registry: None,
#[cfg(feature = "enterprise")]
extension_range_provider_factory: None,
}
}
@@ -199,6 +203,15 @@ impl DatanodeBuilder {
self
}
#[cfg(feature = "enterprise")]
pub fn with_extension_range_provider(
&mut self,
extension_range_provider_factory: mito2::extension::BoxedExtensionRangeProviderFactory,
) -> &mut Self {
self.extension_range_provider_factory = Some(extension_range_provider_factory);
self
}
pub async fn build(mut self) -> Result<Datanode> {
let node_id = self.opts.node_id.context(MissingNodeIdSnafu)?;
@@ -340,7 +353,7 @@ impl DatanodeBuilder {
}
async fn new_region_server(
&self,
&mut self,
schema_metadata_manager: SchemaMetadataManagerRef,
event_listener: RegionServerEventListenerRef,
) -> Result<RegionServer> {
@@ -376,13 +389,13 @@ impl DatanodeBuilder {
);
let object_store_manager = Self::build_object_store_manager(&opts.storage).await?;
let engines = Self::build_store_engines(
opts,
object_store_manager,
schema_metadata_manager,
self.plugins.clone(),
)
.await?;
let engines = self
.build_store_engines(
object_store_manager,
schema_metadata_manager,
self.plugins.clone(),
)
.await?;
for engine in engines {
region_server.register_engine(engine);
}
@@ -394,7 +407,7 @@ impl DatanodeBuilder {
/// Builds [RegionEngineRef] from `store_engine` section in `opts`
async fn build_store_engines(
opts: &DatanodeOptions,
&mut self,
object_store_manager: ObjectStoreManagerRef,
schema_metadata_manager: SchemaMetadataManagerRef,
plugins: Plugins,
@@ -403,7 +416,7 @@ impl DatanodeBuilder {
let mut mito_engine_config = MitoConfig::default();
let mut file_engine_config = file_engine::config::EngineConfig::default();
for engine in &opts.region_engine {
for engine in &self.opts.region_engine {
match engine {
RegionEngineConfig::Mito(config) => {
mito_engine_config = config.clone();
@@ -417,14 +430,14 @@ impl DatanodeBuilder {
}
}
let mito_engine = Self::build_mito_engine(
opts,
object_store_manager.clone(),
mito_engine_config,
schema_metadata_manager.clone(),
plugins.clone(),
)
.await?;
let mito_engine = self
.build_mito_engine(
object_store_manager.clone(),
mito_engine_config,
schema_metadata_manager.clone(),
plugins.clone(),
)
.await?;
let metric_engine = MetricEngine::try_new(mito_engine.clone(), metric_engine_config)
.context(BuildMetricEngineSnafu)?;
@@ -443,12 +456,13 @@ impl DatanodeBuilder {
/// Builds [MitoEngine] according to options.
async fn build_mito_engine(
opts: &DatanodeOptions,
&mut self,
object_store_manager: ObjectStoreManagerRef,
mut config: MitoConfig,
schema_metadata_manager: SchemaMetadataManagerRef,
plugins: Plugins,
) -> Result<MitoEngine> {
let opts = &self.opts;
if opts.storage.is_object_storage() {
// Enable the write cache when setting object storage
config.enable_write_cache = true;
@@ -456,17 +470,27 @@ impl DatanodeBuilder {
}
let mito_engine = match &opts.wal {
DatanodeWalConfig::RaftEngine(raft_engine_config) => MitoEngine::new(
&opts.storage.data_home,
config,
Self::build_raft_engine_log_store(&opts.storage.data_home, raft_engine_config)
.await?,
object_store_manager,
schema_metadata_manager,
plugins,
)
.await
.context(BuildMitoEngineSnafu)?,
DatanodeWalConfig::RaftEngine(raft_engine_config) => {
let log_store =
Self::build_raft_engine_log_store(&opts.storage.data_home, raft_engine_config)
.await?;
let builder = MitoEngineBuilder::new(
&opts.storage.data_home,
config,
log_store,
object_store_manager,
schema_metadata_manager,
plugins,
);
#[cfg(feature = "enterprise")]
let builder = builder.with_extension_range_provider_factory(
self.extension_range_provider_factory.take(),
);
builder.try_build().await.context(BuildMitoEngineSnafu)?
}
DatanodeWalConfig::Kafka(kafka_config) => {
if kafka_config.create_index && opts.node_id.is_none() {
warn!("The WAL index creation only available in distributed mode.")
@@ -488,16 +512,21 @@ impl DatanodeBuilder {
None
};
MitoEngine::new(
let builder = MitoEngineBuilder::new(
&opts.storage.data_home,
config,
Self::build_kafka_log_store(kafka_config, global_index_collector).await?,
object_store_manager,
schema_metadata_manager,
plugins,
)
.await
.context(BuildMitoEngineSnafu)?
);
#[cfg(feature = "enterprise")]
let builder = builder.with_extension_range_provider_factory(
self.extension_range_provider_factory.take(),
);
builder.try_build().await.context(BuildMitoEngineSnafu)?
}
};
Ok(mito_engine)

View File

@@ -142,14 +142,6 @@ pub enum Error {
source: Box<log_store::error::Error>,
},
#[snafu(display("Failed to init backend"))]
InitBackend {
#[snafu(source)]
error: object_store::Error,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Invalid SQL, error: {}", msg))]
InvalidSql { msg: String },
@@ -387,6 +379,29 @@ pub enum Error {
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Failed to serialize json"))]
SerializeJson {
#[snafu(source)]
error: serde_json::Error,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Failed object store operation"))]
ObjectStore {
source: object_store::error::Error,
#[snafu(implicit)]
location: Location,
},
#[snafu(display("Failed to build cache store"))]
BuildCacheStore {
#[snafu(source)]
error: object_store::Error,
#[snafu(implicit)]
location: Location,
},
}
pub type Result<T> = std::result::Result<T, Error>;
@@ -439,8 +454,6 @@ impl ErrorExt for Error {
StartServer { source, .. } | ShutdownServer { source, .. } => source.status_code(),
InitBackend { .. } => StatusCode::StorageUnavailable,
OpenLogStore { source, .. } => source.status_code(),
MetaClientInit { source, .. } => source.status_code(),
UnsupportedOutput { .. } => StatusCode::Unsupported,
@@ -457,6 +470,10 @@ impl ErrorExt for Error {
StatusCode::RegionBusy
}
MissingCache { .. } => StatusCode::Internal,
SerializeJson { .. } => StatusCode::Internal,
ObjectStore { source, .. } => source.status_code(),
BuildCacheStore { .. } => StatusCode::StorageUnavailable,
}
}

View File

@@ -278,7 +278,7 @@ mod tests {
let mut region_server = mock_region_server();
let heartbeat_handler = RegionHeartbeatResponseHandler::new(region_server.clone());
let mut engine_env = TestEnv::with_prefix("close-region");
let mut engine_env = TestEnv::with_prefix("close-region").await;
let engine = engine_env.create_engine(MitoConfig::default()).await;
region_server.register_engine(Arc::new(engine));
let region_id = RegionId::new(1024, 1);
@@ -326,7 +326,7 @@ mod tests {
let mut region_server = mock_region_server();
let heartbeat_handler = RegionHeartbeatResponseHandler::new(region_server.clone());
let mut engine_env = TestEnv::with_prefix("open-region");
let mut engine_env = TestEnv::with_prefix("open-region").await;
let engine = engine_env.create_engine(MitoConfig::default()).await;
region_server.register_engine(Arc::new(engine));
let region_id = RegionId::new(1024, 1);
@@ -374,7 +374,7 @@ mod tests {
let mut region_server = mock_region_server();
let heartbeat_handler = RegionHeartbeatResponseHandler::new(region_server.clone());
let mut engine_env = TestEnv::with_prefix("open-not-exists-region");
let mut engine_env = TestEnv::with_prefix("open-not-exists-region").await;
let engine = engine_env.create_engine(MitoConfig::default()).await;
region_server.register_engine(Arc::new(engine));
let region_id = RegionId::new(1024, 1);
@@ -406,7 +406,7 @@ mod tests {
let mut region_server = mock_region_server();
let heartbeat_handler = RegionHeartbeatResponseHandler::new(region_server.clone());
let mut engine_env = TestEnv::with_prefix("downgrade-region");
let mut engine_env = TestEnv::with_prefix("downgrade-region").await;
let engine = engine_env.create_engine(MitoConfig::default()).await;
region_server.register_engine(Arc::new(engine));
let region_id = RegionId::new(1024, 1);

View File

@@ -20,12 +20,14 @@ use std::time::Duration;
use api::region::RegionResponse;
use api::v1::region::sync_request::ManifestInfo;
use api::v1::region::{region_request, RegionResponse as RegionResponseV1, SyncRequest};
use api::v1::region::{
region_request, ListMetadataRequest, RegionResponse as RegionResponseV1, SyncRequest,
};
use api::v1::{ResponseHeader, Status};
use arrow_flight::{FlightData, Ticket};
use async_trait::async_trait;
use bytes::Bytes;
use common_error::ext::BoxedError;
use common_error::ext::{BoxedError, ErrorExt};
use common_error::status_code::StatusCode;
use common_query::request::QueryRequest;
use common_query::OutputData;
@@ -47,6 +49,7 @@ pub use query::dummy_catalog::{
DummyCatalogList, DummyTableProviderFactory, TableProviderFactoryRef,
};
use query::QueryEngineRef;
use serde_json;
use servers::error::{self as servers_error, ExecuteGrpcRequestSnafu, Result as ServerResult};
use servers::grpc::flight::{FlightCraft, FlightRecordBatchStream, TonicStream};
use servers::grpc::region_server::RegionServerHandler;
@@ -71,10 +74,10 @@ use tonic::{Request, Response, Result as TonicResult};
use crate::error::{
self, BuildRegionRequestsSnafu, ConcurrentQueryLimiterClosedSnafu,
ConcurrentQueryLimiterTimeoutSnafu, DataFusionSnafu, DecodeLogicalPlanSnafu,
ExecuteLogicalPlanSnafu, FindLogicalRegionsSnafu, HandleBatchDdlRequestSnafu,
HandleBatchOpenRequestSnafu, HandleRegionRequestSnafu, NewPlanDecoderSnafu,
RegionEngineNotFoundSnafu, RegionNotFoundSnafu, RegionNotReadySnafu, Result,
StopRegionEngineSnafu, UnexpectedSnafu, UnsupportedOutputSnafu,
ExecuteLogicalPlanSnafu, FindLogicalRegionsSnafu, GetRegionMetadataSnafu,
HandleBatchDdlRequestSnafu, HandleBatchOpenRequestSnafu, HandleRegionRequestSnafu,
NewPlanDecoderSnafu, RegionEngineNotFoundSnafu, RegionNotFoundSnafu, RegionNotReadySnafu,
Result, SerializeJsonSnafu, StopRegionEngineSnafu, UnexpectedSnafu, UnsupportedOutputSnafu,
};
use crate::event_listener::RegionServerEventListenerRef;
@@ -138,12 +141,12 @@ impl RegionServer {
/// Finds the region's engine by its id. If the region is not ready, returns `None`.
pub fn find_engine(&self, region_id: RegionId) -> Result<Option<RegionEngineRef>> {
self.inner
.get_engine(region_id, &RegionChange::None)
.map(|x| match x {
CurrentEngine::Engine(engine) => Some(engine),
CurrentEngine::EarlyReturn(_) => None,
})
match self.inner.get_engine(region_id, &RegionChange::None) {
Ok(CurrentEngine::Engine(engine)) => Ok(Some(engine)),
Ok(CurrentEngine::EarlyReturn(_)) => Ok(None),
Err(error::Error::RegionNotFound { .. }) => Ok(None),
Err(err) => Err(err),
}
}
#[tracing::instrument(skip_all)]
@@ -412,6 +415,7 @@ impl RegionServer {
Ok(RegionResponse {
affected_rows,
extensions,
metadata: Vec::new(),
})
}
@@ -441,6 +445,7 @@ impl RegionServer {
Ok(RegionResponse {
affected_rows,
extensions,
metadata: Vec::new(),
})
}
@@ -473,6 +478,48 @@ impl RegionServer {
.map(|_| RegionResponse::new(AffectedRows::default()))
}
/// Handles the ListMetadata request and retrieves metadata for specified regions.
///
/// Returns the results as a JSON-serialized list in the [RegionResponse]. It serializes
/// non-existing regions as `null`.
#[tracing::instrument(skip_all)]
async fn handle_list_metadata_request(
&self,
request: &ListMetadataRequest,
) -> Result<RegionResponse> {
let mut region_metadatas = Vec::new();
// Collect metadata for each region
for region_id in &request.region_ids {
let region_id = RegionId::from_u64(*region_id);
// Get the engine.
let Some(engine) = self.find_engine(region_id)? else {
region_metadatas.push(None);
continue;
};
match engine.get_metadata(region_id).await {
Ok(metadata) => region_metadatas.push(Some(metadata)),
Err(err) => {
if err.status_code() == StatusCode::RegionNotFound {
region_metadatas.push(None);
} else {
Err(err).with_context(|_| GetRegionMetadataSnafu {
engine: engine.name(),
region_id,
})?;
}
}
}
}
// Serialize metadata to JSON
let json_result = serde_json::to_vec(&region_metadatas).context(SerializeJsonSnafu)?;
let response = RegionResponse::from_metadata(json_result);
Ok(response)
}
/// Sync region manifest and registers new opened logical regions.
pub async fn sync_region(
&self,
@@ -504,6 +551,10 @@ impl RegionServerHandler for RegionServer {
region_request::Body::Sync(sync_request) => {
self.handle_sync_region_request(sync_request).await
}
region_request::Body::ListMetadata(list_metadata_request) => {
self.handle_list_metadata_request(list_metadata_request)
.await
}
_ => self.handle_requests_in_serial(request).await,
}
.map_err(BoxedError::new)
@@ -518,6 +569,7 @@ impl RegionServerHandler for RegionServer {
}),
affected_rows: response.affected_rows as _,
extensions: response.extensions,
metadata: response.metadata,
})
}
}
@@ -897,6 +949,7 @@ impl RegionServerInner {
Ok(RegionResponse {
affected_rows: result.affected_rows,
extensions: result.extensions,
metadata: Vec::new(),
})
}
Err(err) => {
@@ -967,6 +1020,7 @@ impl RegionServerInner {
Ok(RegionResponse {
affected_rows: result.affected_rows,
extensions: result.extensions,
metadata: Vec::new(),
})
}
Err(err) => {
@@ -1242,8 +1296,11 @@ mod tests {
use std::assert_matches::assert_matches;
use api::v1::SemanticType;
use common_error::ext::ErrorExt;
use datatypes::prelude::ConcreteDataType;
use mito2::test_util::CreateRequestBuilder;
use store_api::metadata::{ColumnMetadata, RegionMetadata, RegionMetadataBuilder};
use store_api::region_engine::RegionEngine;
use store_api::region_request::{RegionDropRequest, RegionOpenRequest, RegionTruncateRequest};
use store_api::storage::RegionId;
@@ -1605,4 +1662,175 @@ mod tests {
let forth_query = p.acquire().await;
assert!(forth_query.is_ok());
}
fn mock_region_metadata(region_id: RegionId) -> RegionMetadata {
let mut metadata_builder = RegionMetadataBuilder::new(region_id);
metadata_builder.push_column_metadata(ColumnMetadata {
column_schema: datatypes::schema::ColumnSchema::new(
"timestamp",
ConcreteDataType::timestamp_nanosecond_datatype(),
false,
),
semantic_type: SemanticType::Timestamp,
column_id: 0,
});
metadata_builder.push_column_metadata(ColumnMetadata {
column_schema: datatypes::schema::ColumnSchema::new(
"file",
ConcreteDataType::string_datatype(),
true,
),
semantic_type: SemanticType::Tag,
column_id: 1,
});
metadata_builder.push_column_metadata(ColumnMetadata {
column_schema: datatypes::schema::ColumnSchema::new(
"message",
ConcreteDataType::string_datatype(),
true,
),
semantic_type: SemanticType::Field,
column_id: 2,
});
metadata_builder.primary_key(vec![1]);
metadata_builder.build().unwrap()
}
#[tokio::test]
async fn test_handle_list_metadata_request() {
common_telemetry::init_default_ut_logging();
let mut mock_region_server = mock_region_server();
let region_id_1 = RegionId::new(1, 0);
let region_id_2 = RegionId::new(2, 0);
let metadata_1 = mock_region_metadata(region_id_1);
let metadata_2 = mock_region_metadata(region_id_2);
let metadatas = vec![Some(metadata_1.clone()), Some(metadata_2.clone())];
let metadata_1 = Arc::new(metadata_1);
let metadata_2 = Arc::new(metadata_2);
let (engine, _) = MockRegionEngine::with_metadata_mock_fn(
MITO_ENGINE_NAME,
Box::new(move |region_id| {
if region_id == region_id_1 {
Ok(metadata_1.clone())
} else if region_id == region_id_2 {
Ok(metadata_2.clone())
} else {
error::RegionNotFoundSnafu { region_id }.fail()
}
}),
);
mock_region_server.register_engine(engine.clone());
mock_region_server
.inner
.region_map
.insert(region_id_1, RegionEngineWithStatus::Ready(engine.clone()));
mock_region_server
.inner
.region_map
.insert(region_id_2, RegionEngineWithStatus::Ready(engine.clone()));
// All regions exist.
let list_metadata_request = ListMetadataRequest {
region_ids: vec![region_id_1.as_u64(), region_id_2.as_u64()],
};
let response = mock_region_server
.handle_list_metadata_request(&list_metadata_request)
.await
.unwrap();
let decoded_metadata: Vec<Option<RegionMetadata>> =
serde_json::from_slice(&response.metadata).unwrap();
assert_eq!(metadatas, decoded_metadata);
}
#[tokio::test]
async fn test_handle_list_metadata_not_found() {
common_telemetry::init_default_ut_logging();
let mut mock_region_server = mock_region_server();
let region_id_1 = RegionId::new(1, 0);
let region_id_2 = RegionId::new(2, 0);
let metadata_1 = mock_region_metadata(region_id_1);
let metadatas = vec![Some(metadata_1.clone()), None];
let metadata_1 = Arc::new(metadata_1);
let (engine, _) = MockRegionEngine::with_metadata_mock_fn(
MITO_ENGINE_NAME,
Box::new(move |region_id| {
if region_id == region_id_1 {
Ok(metadata_1.clone())
} else {
error::RegionNotFoundSnafu { region_id }.fail()
}
}),
);
mock_region_server.register_engine(engine.clone());
mock_region_server
.inner
.region_map
.insert(region_id_1, RegionEngineWithStatus::Ready(engine.clone()));
// Not in region map.
let list_metadata_request = ListMetadataRequest {
region_ids: vec![region_id_1.as_u64(), region_id_2.as_u64()],
};
let response = mock_region_server
.handle_list_metadata_request(&list_metadata_request)
.await
.unwrap();
let decoded_metadata: Vec<Option<RegionMetadata>> =
serde_json::from_slice(&response.metadata).unwrap();
assert_eq!(metadatas, decoded_metadata);
// Not in region engine.
mock_region_server
.inner
.region_map
.insert(region_id_2, RegionEngineWithStatus::Ready(engine.clone()));
let response = mock_region_server
.handle_list_metadata_request(&list_metadata_request)
.await
.unwrap();
let decoded_metadata: Vec<Option<RegionMetadata>> =
serde_json::from_slice(&response.metadata).unwrap();
assert_eq!(metadatas, decoded_metadata);
}
#[tokio::test]
async fn test_handle_list_metadata_failed() {
common_telemetry::init_default_ut_logging();
let mut mock_region_server = mock_region_server();
let region_id_1 = RegionId::new(1, 0);
let (engine, _) = MockRegionEngine::with_metadata_mock_fn(
MITO_ENGINE_NAME,
Box::new(move |region_id| {
error::UnexpectedSnafu {
violated: format!("Failed to get region {region_id}"),
}
.fail()
}),
);
mock_region_server.register_engine(engine.clone());
mock_region_server
.inner
.region_map
.insert(region_id_1, RegionEngineWithStatus::Ready(engine.clone()));
// Failed to get.
let list_metadata_request = ListMetadataRequest {
region_ids: vec![region_id_1.as_u64()],
};
mock_region_server
.handle_list_metadata_request(&list_metadata_request)
.await
.unwrap_err();
}
}

View File

@@ -14,45 +14,22 @@
//! object storage utilities
mod azblob;
pub mod fs;
mod gcs;
mod oss;
mod s3;
use std::path;
use std::path::Path;
use std::sync::Arc;
use std::time::Duration;
use common_telemetry::{info, warn};
use mito2::access_layer::{ATOMIC_WRITE_DIR, OLD_ATOMIC_WRITE_DIR};
use object_store::factory::new_raw_object_store;
use object_store::layers::{LruCacheLayer, RetryInterceptor, RetryLayer};
use object_store::services::Fs;
use object_store::util::{join_dir, normalize_dir, with_instrument_layers};
use object_store::{Access, Error, HttpClient, ObjectStore, ObjectStoreBuilder};
use object_store::util::{clean_temp_dir, join_dir, with_instrument_layers};
use object_store::{
Access, Error, ObjectStore, ObjectStoreBuilder, ATOMIC_WRITE_DIR, OLD_ATOMIC_WRITE_DIR,
};
use snafu::prelude::*;
use crate::config::{HttpClientConfig, ObjectStoreConfig, DEFAULT_OBJECT_STORE_CACHE_SIZE};
use crate::error::{self, BuildHttpClientSnafu, CreateDirSnafu, Result};
pub(crate) async fn new_raw_object_store(
store: &ObjectStoreConfig,
data_home: &str,
) -> Result<ObjectStore> {
let data_home = normalize_dir(data_home);
let object_store = match store {
ObjectStoreConfig::File(file_config) => {
fs::new_fs_object_store(&data_home, file_config).await
}
ObjectStoreConfig::S3(s3_config) => s3::new_s3_object_store(s3_config).await,
ObjectStoreConfig::Oss(oss_config) => oss::new_oss_object_store(oss_config).await,
ObjectStoreConfig::Azblob(azblob_config) => {
azblob::new_azblob_object_store(azblob_config).await
}
ObjectStoreConfig::Gcs(gcs_config) => gcs::new_gcs_object_store(gcs_config).await,
}?;
Ok(object_store)
}
use crate::config::{ObjectStoreConfig, DEFAULT_OBJECT_STORE_CACHE_SIZE};
use crate::error::{self, CreateDirSnafu, Result};
fn with_retry_layers(object_store: ObjectStore) -> ObjectStore {
object_store.layer(
@@ -66,7 +43,9 @@ pub(crate) async fn new_object_store_without_cache(
store: &ObjectStoreConfig,
data_home: &str,
) -> Result<ObjectStore> {
let object_store = new_raw_object_store(store, data_home).await?;
let object_store = new_raw_object_store(store, data_home)
.await
.context(error::ObjectStoreSnafu)?;
// Enable retry layer and cache layer for non-fs object storages
let object_store = if store.is_object_storage() {
// Adds retry layer
@@ -83,7 +62,9 @@ pub(crate) async fn new_object_store(
store: ObjectStoreConfig,
data_home: &str,
) -> Result<ObjectStore> {
let object_store = new_raw_object_store(&store, data_home).await?;
let object_store = new_raw_object_store(&store, data_home)
.await
.context(error::ObjectStoreSnafu)?;
// Enable retry layer and cache layer for non-fs object storages
let object_store = if store.is_object_storage() {
let object_store = if let Some(cache_layer) = build_cache_layer(&store, data_home).await? {
@@ -170,20 +151,20 @@ async fn build_cache_layer(
&& !path.trim().is_empty()
{
let atomic_temp_dir = join_dir(path, ATOMIC_WRITE_DIR);
clean_temp_dir(&atomic_temp_dir)?;
clean_temp_dir(&atomic_temp_dir).context(error::ObjectStoreSnafu)?;
// Compatible code. Remove this after a major release.
let old_atomic_temp_dir = join_dir(path, OLD_ATOMIC_WRITE_DIR);
clean_temp_dir(&old_atomic_temp_dir)?;
clean_temp_dir(&old_atomic_temp_dir).context(error::ObjectStoreSnafu)?;
let cache_store = Fs::default()
.root(path)
.atomic_write_dir(&atomic_temp_dir)
.build()
.context(error::InitBackendSnafu)?;
.context(error::BuildCacheStoreSnafu)?;
let cache_layer = LruCacheLayer::new(Arc::new(cache_store), cache_capacity.0 as usize)
.context(error::InitBackendSnafu)?;
.context(error::BuildCacheStoreSnafu)?;
cache_layer.recover_cache(false).await;
info!(
"Enabled local object storage cache, path: {}, capacity: {}.",
@@ -196,31 +177,6 @@ async fn build_cache_layer(
}
}
pub(crate) fn clean_temp_dir(dir: &str) -> Result<()> {
if path::Path::new(&dir).exists() {
info!("Begin to clean temp storage directory: {}", dir);
std::fs::remove_dir_all(dir).context(error::RemoveDirSnafu { dir })?;
info!("Cleaned temp storage directory: {}", dir);
}
Ok(())
}
pub(crate) fn build_http_client(config: &HttpClientConfig) -> Result<HttpClient> {
if config.skip_ssl_validation {
common_telemetry::warn!("Skipping SSL validation for object storage HTTP client. Please ensure the environment is trusted.");
}
let client = reqwest::ClientBuilder::new()
.pool_max_idle_per_host(config.pool_max_idle_per_host as usize)
.connect_timeout(config.connect_timeout)
.pool_idle_timeout(config.pool_idle_timeout)
.timeout(config.timeout)
.danger_accept_invalid_certs(config.skip_ssl_validation)
.build()
.context(BuildHttpClientSnafu)?;
Ok(HttpClient::with(client))
}
struct PrintDetailedError;
// PrintDetailedError is a retry interceptor that prints error in Debug format in retrying.

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