Commit Graph

  • 7fb4595c7e fix: WeakHandle was holding on to the Timeline allocation Christian Schwarz 2025-01-15 20:44:20 +01:00
  • 350dc251df test case demonstrates the issue: we hod Timeline object alive Christian Schwarz 2025-01-15 21:44:28 +01:00
  • 045b05cd1b Merge pull request #10408 from neondatabase/rc/release-compute/2025-01-15 release-compute-7552 Tristan Partin 2025-01-15 14:27:34 -06:00
  • fb0e2acb2f pageserver: add page_trace API for debugging (#10293) John Spray 2025-01-15 19:07:22 +00:00
  • 14318afcc0 Merge branch '22100-change-fastimport-db-name' into 22037-basic-fast-import-e2e Gleb Novikov 2025-01-15 19:02:03 +00:00
  • ebc4735bf4 postgres waiting timeout & retry as constants Gleb Novikov 2025-01-15 17:45:21 +00:00
  • 5b77a6d3ce address clippy Christian Schwarz 2025-01-15 19:38:21 +01:00
  • 8c5005ff59 rename IoConcurrency::{todo=>serial} and remove deprecation warning Christian Schwarz 2025-01-15 19:38:05 +01:00
  • f8218ac5fc Revert "investigation: add log_if_slow => shows that the io_futures are slow" Christian Schwarz 2025-01-15 19:34:37 +01:00
  • 40470c66cd remove opportunistic poll, it seems slightly beneficial for perf Christian Schwarz 2025-01-15 19:16:55 +01:00
  • 9b9479881a extend script with instructions to configure batching Christian Schwarz 2025-01-15 19:27:48 +01:00
  • efaec6cdf8 Add endpoint and storcon cli cmd to set sk scheduling policy (#10400) Arpad Müller 2025-01-15 19:15:30 +01:00
  • af11b201bd now the issue is no longer reproducible, maybe it was the barriers? Christian Schwarz 2025-01-15 19:10:45 +01:00
  • 8fafff37c5 remove the whole barriers business Christian Schwarz 2025-01-15 19:00:00 +01:00
  • 6a4d8ec410 Compute release 2025-01-15 github-actions[bot] 2025-01-15 17:59:13 +00:00
  • e81fa7137e investigation: add log_if_slow => shows that the io_futures are slow Christian Schwarz 2025-01-15 18:56:07 +01:00
  • e60738f029 it's reproducible before the merge, so, continuing to investigate and fix here Christian Schwarz 2025-01-15 18:43:01 +01:00
  • f75b07a160 I find that if I ever go beyond queue-depth=4, something in the pageserver locks up. Christian Schwarz 2025-01-15 18:31:40 +01:00
  • a5524fcf4d add comment to use queue-depthed pagebench to the script Christian Schwarz 2025-01-15 18:31:29 +01:00
  • 4c2ee6a011 added 10 min timeout on waiting loop Gleb Novikov 2025-01-15 17:15:58 +00:00
  • c09d817c98 review comments Gleb Novikov 2025-01-15 17:13:52 +00:00
  • 3d41069dc4 Update pgrx in extension builds to 0.12.9 (#10372) Tristan Partin 2025-01-15 10:26:58 -06:00
  • 351da2349e Merge branch 'problame/hung-shutdown/fix' into vlad/read-path-concurrent-io Christian Schwarz 2025-01-15 17:09:02 +01:00
  • dbebede7bf safekeeper: fan out from single wal reader to multiple shards (#10190) Vlad Lazar 2025-01-15 15:33:54 +00:00
  • 3e529f124f Remove leading slashes when downloading remote files (#10396) Tristan Partin 2025-01-15 09:29:52 -06:00
  • db5513076a Add PAMessageSerialize and ProposerAcceptorGreeting v3 Arseny Sher 2025-01-14 18:29:03 +01:00
  • 70d4e077a6 pgindent wp code Arseny Sher 2024-12-26 08:53:17 +01:00
  • ae9db8975a Add START_WAL_PUSH proto_version and allow_timeline_creation options. Arseny Sher 2025-01-15 12:35:13 +01:00
  • c545d227b9 review doc comment Christian Schwarz 2025-01-15 16:24:39 +01:00
  • a4fc6a92c9 fix cargo doc Christian Schwarz 2025-01-15 16:10:04 +01:00
  • 05a71c7d6a safekeeper: add membership configuration switch endpoint (#10241) Arseny Sher 2025-01-15 17:16:04 +03:00
  • 2205736262 doc comment & one fixup Christian Schwarz 2025-01-15 14:24:29 +01:00
  • b9464865b6 benchmarks: report successful runs to slack as well (#10393) Alexander Bayandin 2025-01-15 13:05:05 +00:00
  • 1577430408 safekeeper: decode and interpret for multiple shards in one go (#10201) Vlad Lazar 2025-01-15 11:10:24 +00:00
  • 05d17a10ae rfc: add CPU and heap profiling RFC (#10085) Erik Grinaker 2025-01-15 11:35:38 +01:00
  • 2d0ea08524 Add safekeeper membership conf to control file. (#10196) Arseny Sher 2025-01-15 12:45:58 +03:00
  • c98cbbeac1 Add migration details to safekeeper membership RFC. (#10272) Arseny Sher 2025-01-15 12:41:49 +03:00
  • c4d42d1b2e fix tests conrad/another-jwt-auth-demo Conrad Ludgate 2025-01-15 09:33:36 +00:00
  • 5f9ddbae2f Merge branch 'problame/hung-shutdown/demo-hypothesis' into problame/hung-shutdown/fix Christian Schwarz 2025-01-15 00:25:11 +01:00
  • 173f18832c fixup problame/hung-shutdown/demo-hypothesis Christian Schwarz 2025-01-15 00:24:59 +01:00
  • 23bd5833e1 Merge branch 'problame/hung-shutdown/demo-hypothesis' into problame/hung-shutdown/fix Christian Schwarz 2025-01-15 00:21:54 +01:00
  • dedd524d7e refinements Christian Schwarz 2025-01-15 00:21:28 +01:00
  • 0340f00228 post-merge fix the handling of the new pagestream Test message, so that the regression test now passes Christian Schwarz 2025-01-14 23:56:35 +01:00
  • 366ff9ffcc Merge branch 'problame/hung-shutdown/demo-hypothesis' into problame/hung-shutdown/fix Christian Schwarz 2025-01-14 23:51:53 +01:00
  • a8f9b564be fix cd pageserver && cargo clippy --features testing build Christian Schwarz 2025-01-14 23:50:22 +01:00
  • 5450e54dab bump ci Christian Schwarz 2025-01-14 22:47:16 +01:00
  • 53b05c4ba0 cleanups to make CI pass (well, fail because the bug isn't fixed yet) Christian Schwarz 2025-01-14 22:45:09 +01:00
  • 47c1640acc storage controller: pagination for tenant listing API (#10365) John Spray 2025-01-14 21:37:32 +00:00
  • 1f7d173235 Merge remote-tracking branch 'origin/main' into problame/hung-shutdown/demo-hypothesis Christian Schwarz 2025-01-14 22:33:20 +01:00
  • 8454e19a0f address warnings and such Christian Schwarz 2025-01-14 22:28:08 +01:00
  • 45e08d0aa5 it repros Christian Schwarz 2025-01-14 22:16:27 +01:00
  • 6debb49b87 pageserver: coalesce index uploads when possible (#10248) Erik Grinaker 2025-01-14 22:10:17 +01:00
  • 9a02bc0cfd try to repro root cause hypothesis for https://github.com/neondatabase/neon/issues/10309 Christian Schwarz 2025-01-14 20:42:01 +01:00
  • 7dc86ec883 pageserver: API for invoking page trace jcsp/pagetrace-releasebased John Spray 2025-01-07 11:52:04 +00:00
  • 44f1c4917d pageserver: add PageTrace machinery John Spray 2025-01-07 11:31:33 +00:00
  • e58e29e639 pageserver: limit number of upload queue tasks (#10384) Erik Grinaker 2025-01-14 19:01:14 +01:00
  • d36112d20f Simplify compute dockerfile by setting PATH just once (#10357) Heikki Linnakangas 2025-01-14 19:02:35 +02:00
  • eab0be6fa8 feat: allow multiple IP patterns in JWT Conrad Ludgate 2025-01-14 16:53:36 +00:00
  • ffaa52ff5d pageserver: reorder upload queue when possible (#10218) Erik Grinaker 2025-01-14 17:31:59 +01:00
  • aa7323a384 storage controller: quality of life improvements for AZ handling (#10379) John Spray 2025-01-14 15:30:43 +00:00
  • 2466a2f977 page_service: throttle individual requests instead of the batched request (#10353) Christian Schwarz 2025-01-14 16:28:01 +01:00
  • 9bdb14c1c0 fix(pageserver): ensure initial image layers have correct key ranges (#10374) Alex Chi Z. 2025-01-14 10:27:48 -05:00
  • 6c0cf1c1cc fix matrix BodoBolero 2025-01-14 16:13:06 +01:00
  • 95d50f918a syntax error BodoBolero 2025-01-14 16:10:44 +01:00
  • 30ba4ac50b test different stripe sizes BodoBolero 2025-01-14 16:09:08 +01:00
  • 9e1cd986d7 address "why mut" nits; https://github.com/neondatabase/neon/pull/10353#discussion_r1913685991 https://github.com/neondatabase/neon/pull/10353#discussion_r1913683065 https://github.com/neondatabase/neon/pull/10353#discussion_r1913683392 Christian Schwarz 2025-01-14 15:30:33 +01:00
  • 47544dcc0b simplify SmgrOpTimerState variant names, and add some doc comments; https://github.com/neondatabase/neon/pull/10353#discussion_r1913676569 and https://github.com/neondatabase/neon/pull/10353#discussion_r1913676824 Christian Schwarz 2025-01-14 15:28:09 +01:00
  • 4e094d9638 rearrange code & inline HandleInner::shutdown() to minimize the diff Christian Schwarz 2025-01-14 15:12:46 +01:00
  • c8bee86586 in some early WIP commit we had removed the loop{} inside get(); re-establish it one level down Christian Schwarz 2025-01-14 15:12:03 +01:00
  • 768a867dcf doc comment fix Christian Schwarz 2025-01-14 14:54:15 +01:00
  • 3b65465e10 turns out with the switch to sync Mutex there's no reason for upgrade() to be async either Christian Schwarz 2025-01-14 14:53:41 +01:00
  • e4ea706424 turns out PerTimelineState::shutdown() doesn't need to be async Christian Schwarz 2025-01-14 14:48:03 +01:00
  • 7034f54a9e remove the earlier-commented-out assertions on arc reference counts, they were too whiteboxy to begin with Christian Schwarz 2025-01-14 14:35:43 +01:00
  • d68c5ddf7e avoid the tokio::sync::Mutex by wrapping the GateGuard into an Arc Christian Schwarz 2025-01-14 14:23:28 +01:00
  • df4abd8b14 fix: force-refresh azure identity token (#10378) Conrad Ludgate 2025-01-14 12:53:32 +00:00
  • f64a240888 fmt Conrad Ludgate 2025-01-14 12:32:50 +00:00
  • b95365b45d Revert "experiment: what if we make Handle !Send so it can't be held across await points" Christian Schwarz 2025-01-14 13:32:12 +01:00
  • b44070d0c7 experiment: what if we make Handle !Send so it can't be held across await points Christian Schwarz 2025-01-14 13:30:48 +01:00
  • 97d0147ed9 validate jwt during auth_quirks Conrad Ludgate 2025-01-14 11:59:45 +00:00
  • a80dcfa544 Capture LD_LIBRARY_PATH from pytest env Gleb Novikov 2025-01-06 17:57:08 +00:00
  • 337ad52c37 Fixed initdb locale Gleb Novikov 2025-01-03 16:56:00 +00:00
  • fd0acb6195 poetry run ruff check --fix . Gleb Novikov 2025-01-03 15:27:25 +00:00
  • a7f8b9f6b5 poetry run ruff format . Gleb Novikov 2025-01-03 15:13:19 +00:00
  • 1e9707f7ee Added todo on full import test with pageserver Gleb Novikov 2025-01-03 15:10:25 +00:00
  • 6d297857a7 Moved test_fast_import to test_import_pgdata Gleb Novikov 2025-01-03 14:54:07 +00:00
  • 6e5a0add43 Implemented basic test of fast import Gleb Novikov 2025-01-03 14:41:49 +00:00
  • e291fb7edc Fixture for fast_import binary is working Gleb Novikov 2025-01-02 18:06:12 +00:00
  • ebe26e218b cargo fmt --all Gleb Novikov 2025-01-02 13:45:20 +00:00
  • 2f0a127e0c Create neondb database and restore into it Gleb Novikov 2024-12-31 16:34:33 +00:00
  • 131ab74be8 effective_io_concurrency=0 on macos Gleb Novikov 2024-12-31 16:34:17 +00:00
  • 3a66eebf4e Made fast_import testable locally (made s3 prefix optional, added source_connection_string param) Gleb Novikov 2024-12-31 15:36:25 +00:00
  • 6b22acba9b avoid cloning the Arc<Timeline> on every handle upgrade/downgade, by wrapping it in yet another Arc Christian Schwarz 2025-01-14 12:28:37 +01:00
  • d4eedb4069 feat: introduce jwks settings in cplane response Conrad Ludgate 2025-01-14 11:26:15 +00:00
  • 22058d17d1 it turns out PerTimelineState need not store a Types::Timeline at all Christian Schwarz 2025-01-14 12:22:00 +01:00
  • a8d096b72c Revert "WIP experiment: avoid upgrading" Christian Schwarz 2025-01-14 12:18:02 +01:00
  • f6eb6fff9f WIP experiment: avoid upgrading Christian Schwarz 2025-01-14 12:14:31 +01:00
  • c868ceded0 WeakHandle should store weak ref to the GateGuard Christian Schwarz 2025-01-14 12:08:53 +01:00
  • e82aa9419e convert handles to named structs Christian Schwarz 2025-01-14 12:01:00 +01:00
  • bfe9efefb8 test tenant::timeline::handle::tests::test_timeline_shutdown hangs Christian Schwarz 2025-01-14 11:56:08 +01:00
  • 62f63275b2 fix test_connection_handler_exit Christian Schwarz 2025-01-14 11:55:45 +01:00