Commit Graph

72 Commits

Author SHA1 Message Date
Heikki Linnakangas e94a5ce360 Rename pg_control_ffi.h to bindgen_deps.h, for clarity.
The pg_control_ffi.h name implies that it only includes stuff related to
pg_control.h. That's mostly true currently, but really the point of the
file is to include everything that we need to generate Rust definitions
from.
2022-08-16 19:37:36 +03:00
Kirill Bulatov 648e8bbefe Fix 1.63 clippy lints (#2282) 2022-08-16 18:49:22 +03:00
Arseny Sher 431393e361 Find end of WAL on safekeepers using WalStreamDecoder.
We could make it inside wal_storage.rs, but taking into account that
 - wal_storage.rs reading is async
 - we don't need s3 here
 - error handling is different; error during decoding is normal
I decided to put it separately.

Test
cargo test test_find_end_of_wal_last_crossing_segment
prepared earlier by @yeputons passes now.

Fixes https://github.com/neondatabase/neon/issues/544
      https://github.com/neondatabase/cloud/issues/2004
Supersedes https://github.com/neondatabase/neon/pull/2066
2022-08-14 14:47:14 +03:00
Egor Suvorov a7bf60631f postgres_ffi/waldecoder: introduce explicit enum State
Previously it was emulated with a combination of nullable fields.
This change should make the logic more readable.
2022-08-12 11:40:46 +03:00
Egor Suvorov 07bb7a2afe postgres_ffi/waldecoder: remove unused startlsn 2022-08-12 11:40:46 +03:00
Egor Suvorov 142e247e85 postgres_ffi/waldecoder: validate more header fields 2022-08-12 11:40:46 +03:00
Ankur Srivastava 84d1bc06a9 refactor: replace lazy-static with once-cell (#2195)
- Replacing all the occurrences of lazy-static with `once-cell::sync::Lazy`
- fixes #1147

Signed-off-by: Ankur Srivastava <best.ankur@gmail.com>
2022-08-05 19:34:04 +02:00
Arthur Petukhovsky 0a958b0ea1 Check find_end_of_wal errors instead of unwrap 2022-08-04 17:56:19 +03:00
Dmitry Rodionov 5f71aa09d3 support running tests against real s3 implementation without mocking 2022-08-04 11:14:05 +03:00
Thang Pham 6a664629fa Add timeline physical size tracking (#2126)
Ref #1902.

- Track the layered timeline's `physical_size` using `pageserver_current_physical_size` metric when updating the layer map.
- Report the local timeline's `physical_size` in timeline GET APIs.
- Add `include-non-incremental-physical-size` URL flag to also report the local timeline's `physical_size_non_incremental` (similar to `logical_size_non_incremental`)
- Add a `UIntGaugeVec` and `UIntGauge` to represent `u64` prometheus metrics

Co-authored-by: Dmitry Rodionov <dmitry@neon.tech>
2022-07-27 12:36:46 -04:00
Dmitry Ivanov 5f4ccae5c5 [proxy] Add the password hack authentication flow (#2095)
[proxy] Add the `password hack` authentication flow

This lets us authenticate users which can use neither
SNI (due to old libpq) nor connection string `options`
(due to restrictions in other client libraries).

Note: `PasswordHack` will accept passwords which are not
encoded in base64 via the "password" field. The assumption
is that most user passwords will be valid utf-8 strings,
and the rest may still be passed via "password_".
2022-07-25 17:23:10 +03:00
Heikki Linnakangas b4c74c0ecd Clean up unnecessary dependencies.
Just to be tidy.
2022-07-20 16:31:25 +03:00
Thang Pham 160e52ec7e Optimize branch creation (#2101)
Resolves #2054

**Context**: branch creation needs to wait for GC to acquire `gc_cs` lock, which prevents creating new timelines during GC. However, because individual timeline GC iteration also requires `compaction_cs` lock, branch creation may also need to wait for compactions of multiple timelines. This results in large latency when creating a new branch, which we advertised as *"instantly"*.

This PR optimizes the latency of branch creation by separating GC into two phases:
1. Collect GC data (branching points, cutoff LSNs, etc)
2. Perform GC for each timeline

The GC bottleneck comes from step 2, which must wait for compaction of multiple timelines. This PR modifies the branch creation and GC functions to allow GC to hold the GC lock only in step 1. As a result, branch creation doesn't need to wait for compaction to finish but only needs to wait for GC data collection step, which is fast.
2022-07-19 14:56:25 -04:00
Egor Suvorov 94003e1ebc postgres_ffi: test restoring from intermediate LSNs by wal_craft 2022-07-15 19:06:50 +03:00
Egor Suvorov 19ea486cde postgres_ffi/xlog_utils: refactor find_end_of_wal test
* Deduce `last_segment` automatically
* Get rid of local `wal_dir`/`wal_seg_size` variables
* Prepare to test parsing of WAL from multiple specific points, not just the start;
  extract `check_end_of_wal` function to check both partial and non-partial WAL segments.
2022-07-15 19:06:50 +03:00
Alexander Bayandin 07acd6ddde Fix clippy warnings in postgres_ffi/build.rs (#2081) 2022-07-13 14:12:11 +01:00
Alexander Bayandin 61cc562822 Make POSTGRES_INSTALL_DIR configurable for build (#2067) 2022-07-13 09:18:11 +01:00
Egor Suvorov f540f115a3 postgres_ffi/wal_craft: simplify API 2022-07-08 18:30:56 +02:00
Egor Suvorov 0b5b2e8e0b postgres_ffi/wal_craft: extract trait Crafter
Make the intent of the code clearer.
2022-07-08 18:30:56 +02:00
Egor Suvorov 60e5dc10e6 postgres_ffi/wal_generate: use 'craft' instead of 'generate'
It does very fine-tuned byte-to-byte WAL crafting, not a sloppy generation.
Hence 'craft' sounds like a better description.
2022-07-08 18:30:56 +02:00
Egor Suvorov 80b7a3b51a Test what happens when XLOG_SWITCH ends on page boundary, fix #1991 2022-07-08 15:37:26 +02:00
Egor Suvorov 85bda437de postgres_ffi/wal_generate: add last_wal_record_xlog_switch and use it in tests
Fix #1190: WalDecoder did not return correct LSN of the next record after
processing a XLOG_SWITCH record
2022-07-08 15:37:26 +02:00
Egor Suvorov c08fa9d562 postgres_ffi/wal_generate: support generating WAL for an already running Postgres server
* ensure_server_config() function is added to ensure the server does not have background processes
  which intervene with WAL generation
* Rework command line syntax
* Add `print-postgres-config` subcommand which prints the required server configuration
2022-07-08 13:56:37 +02:00
Dmitry Rodionov e1e24336b7 review adjustments, bring back timeline_detach and rename it to timeline_delete 2022-07-07 21:20:04 +03:00
Dmitry Rodionov 4c54e4b37d switch to per-tenant attach/detach
download operations of all timelines for one tenant are now grouped
together so when attach is invoked pageserver downloads all of them
and registers them in a single apply_sync_status_update call so
branches can be used safely with attach/detach
2022-07-07 21:20:04 +03:00
Kirill Bulatov 50821c0a3c Return download stream directly from the remote storage API 2022-07-05 21:45:15 +03:00
Kirill Bulatov 7c49abe7d1 Rework etcd timeline updates and their handling 2022-06-23 09:11:27 +03:00
Kirill Bulatov d8a37452c8 Rename ZenithFeedback (#1912) 2022-06-11 00:44:05 +03:00
Arseny Sher a4d8261390 Save Postgres log in test_find_end_of_wal_* tests. 2022-06-09 19:16:43 +04:00
Kirill Bulatov 8a53472e4f Force etcd broker keys to not to intersect 2022-06-08 11:21:05 +03:00
Dmitry Rodionov 6e26588d17 Allow to customize shutdown condition in PostgresBackend
Use it in PageServerHandler to check per thread shutdown condition
from thread_mgr which takes into account tenants and timelines
2022-06-07 22:11:54 +03:00
KlimentSerafimov fecad1ca34 Resolving issue #1745. Added cluster option for SNI data (#1813)
* Added project option in case SNI data is missing. Resolving issue #1745.

* Added invariant checking for project name: if both sni_data and project_name are available then they should match.
2022-06-06 08:14:41 -04:00
Kirill Bulatov 2623193876 Remove pageserver_connstr from WAL stream logic 2022-06-03 17:30:36 +03:00
Arseny Sher 9e108102b3 Silence etcd safekeeper info key parse errors.
When we subscribe to everything, it is ok to receive not only safekeeper
timeline updates.
2022-06-03 16:12:14 +04:00
Kirill Bulatov 5b06599770 Simplify etcd key regex parsing 2022-06-03 00:23:13 +03:00
Kirill Bulatov 7933804284 Fix and test regex parsing 2022-06-03 00:23:13 +03:00
Kirill Bulatov b0c4ec0594 Log storage sync and etcd events a bit better 2022-06-03 00:23:13 +03:00
Kirill Bulatov e5cb727572 Replace callmemaybe with etcd subscriptions on safekeeper timeline info 2022-06-01 16:07:04 +03:00
Arthur Petukhovsky c3e0b6c839 Implement timeline-based metrics in safekeeper (#1823)
Now there's timelines metrics collector, which goes through all timelines and reports metrics only for active ones
2022-05-31 11:10:50 +03:00
Anastasia Lubennikova 6a867bce6d Rename 'zenith_admin' role to 'cloud_admin' 2022-05-30 11:11:01 +03:00
Anastasia Lubennikova 3accde613d Rename contrib/zenith to contrib/neon. Rename custom GUCs:
- zenith.page_server_connstring -> neon.pageserver_connstring
- zenith.zenith_tenant -> neon.tenantid
- zenith.zenith_timeline -> neon.timelineid
- zenith.max_cluster_size -> neon.max_cluster_size
2022-05-30 11:11:01 +03:00
Kian-Meng Ang f1c51a1267 Fix typos 2022-05-28 14:02:05 +03:00
Arseny Sher 54b75248ff s3 WAL offloading staging review.
- Uncomment accidently `self.keep_alive.abort()` commented line, due to this
  task never finished, which blocked launcher.
- Mess up with initialization one more time, to fix offloader trying to back up
  segment 0. Now we initialize all required LSNs in handle_elected,
  where we learn start LSN for the first time.
- Fix blind attempt to provide safekeeper service file with remote storage
  params.
2022-05-27 14:02:52 +04:00
Arseny Sher 0e1bd57c53 Add WAL offloading to s3 on safekeepers.
Separate task is launched for each timeline and stopped when timeline doesn't
need offloading. Decision who offloads is done through etcd leader election;
currently there is no pre condition for participating, that's a TODO.

neon_local and tests infrastructure for remote storage in safekeepers added,
along with the test itself.

ref #1009

Co-authored-by: Anton Shyrabokau <ahtoxa@Antons-MacBook-Pro.local>
2022-05-27 06:19:23 +04:00
Dmitry Rodionov 5a5737278e add simple metrics for remote storage operations
track number of operations and number of their failures
2022-05-26 01:24:52 +03:00
chaitanya sharma c584d90bb9 initial commit, renamed znodeid to nodeid. 2022-05-25 20:11:26 +03:00
Egor Suvorov 73187bfef1 postgres_ffi: find_end_of_wal_segment: clarify code around xl_crc retrieval
It would be better to not update xl_crc/rec_hdr at all when skipping contrecord,
but I would prefer to keep PR #1574 small.
Better audit of `find_end_of_wal_segment` is coming anyway in #544.
2022-05-21 05:25:17 +02:00
Egor Suvorov 967eb38e81 postgres_ffi: find_end_of_wal_segment: fix contrecord skipping
Also enable corresponding test.
2022-05-21 05:25:17 +02:00
Egor Suvorov a124e44866 postgres_ffi: find_end_of_wal_segment: add lots of trace 2022-05-21 05:25:17 +02:00
Egor Suvorov c4b77084af utils: add const_assert! macro 2022-05-21 05:25:17 +02:00