922 Commits

Author SHA1 Message Date
Patrick Insinger
3e1e934a21 use copy_split in freeze 2021-10-05 13:14:19 -07:00
Patrick Insinger
2ccc8f3e36 OrderedVec copy_prefix 2021-10-05 12:57:58 -07:00
Patrick Insinger
77722b4c9d skip a block when larger lsn is found 2021-10-05 12:52:32 -07:00
Patrick Insinger
7ea2251f02 switch to HashMap for block indexing 2021-10-05 12:46:47 -07:00
Patrick Insinger
896a172d46 OrderedVec - copy_split for segsizes 2021-10-05 11:49:13 -07:00
Patrick Insinger
b7c3f02ed1 page versions 2021-10-05 11:20:40 -07:00
Patrick Insinger
98f58fa2ab use for seqsizes in inmemorylayer 2021-10-05 11:18:29 -07:00
Patrick Insinger
adf4ac0ef7 range fixes 2021-10-05 11:18:29 -07:00
Patrick Insinger
9dd1393f0e page version meta 2021-10-05 11:18:27 -07:00
Patrick Insinger
935984018d ordered-vec and seqsizes 2021-10-05 11:14:09 -07:00
Heikki Linnakangas
d806c3a47e pageserver - serialize PageVersion as it is
Removes the need for PageVersionMeta struct.
2021-10-05 11:07:50 -07:00
Egor Suvorov
05fe39088b Readme updates based on a fresher Ubuntu installation experience (#627) 2021-10-05 19:19:25 +03:00
Egor Suvorov
530d3eaf09 Add more details to pageserver and safekeeper docs (#680) 2021-10-05 19:10:50 +03:00
Egor Suvorov
7e190d72a5 Make pageserver_ prefix for common metric names configurable (#681) 2021-10-05 19:06:44 +03:00
Patrick Insinger
9c936034b6 pageserver - fix newer clippy lints 2021-10-05 00:28:14 -07:00
Kirill Bulatov
5719f13cb2 Rework the relish thread model (#689) 2021-10-05 10:15:56 +03:00
Patrick Insinger
d134a9856e pageserver - introduce RepoHarness for testing 2021-10-04 08:36:35 -07:00
Patrick Insinger
664b99b5ac pageserver - use constant TIMELINE_ID for tests 2021-10-04 08:36:35 -07:00
Arseny Sher
4256231eb7 Enable test_start_compute with safekeepers.
It should work now.
2021-10-04 16:50:46 +03:00
Andrey Taranik
ae27490281 wal_acceptors added to tenant creation tests 2021-10-04 08:58:49 +03:00
Andrey Taranik
fbd8ca2ff4 minor code beautification 2021-10-04 08:58:49 +03:00
Andrey Taranik
ec673a5d67 bulk tenant create test added 2021-10-04 08:58:49 +03:00
Max Sharnoff
7fab38c51e Use threadlocal for walreceiver check (#692) 2021-10-01 15:47:45 -07:00
Max Sharnoff
84f7dcd052 Fix clippy errors on nightly (2021-09-29) (#691)
Most of the changes are for the new if-then-panic lint added in
https://github.com/rust-lang/rust-clippy/pull/7669.
2021-10-01 15:45:42 -07:00
Patrick Insinger
7095a5d551 pageserver - reject and backup future layer files
If a layer file is found with LSN after the disk_consistent_lsn, it is
renamed (to avoid conflicts with new layer files) and a warning is logged.
2021-10-01 11:41:39 -07:00
Patrick Insinger
538c2a2a3e pageserver - store timeline metadata durably
The metadata file is now always 512 bytes. The last 4 bytes are a
crc32c checksum of the previous 508 bytes. Padding zeroes are added
between the serde serialization and the start of the checksum.

A single write call is used, and the file is fsyncd after.
On file creation, the parent directory is fsyncd as well.
2021-10-01 11:41:39 -07:00
Patrick Insinger
62f83869f1 pageserver - fsync image/delta layers
Ensure image and delta layer files are durable.
Also, fsync the parent directory to ensure the directory entries are
durable.
2021-10-01 11:41:39 -07:00
Patrick Insinger
69670b61c4 pageserver - use crashsafe_dir utility
Replace usage of std::fs::create_dir/create_dir_all with crashsafe
equivalents.
2021-10-01 11:41:39 -07:00
Patrick Insinger
0a8aaa2c24 zenith_utils - add crashsafe_dir
Utility for creating directories and directory trees in a crash safe
manor.

Minimizes calls to fsync for trees.
2021-10-01 11:41:39 -07:00
Heikki Linnakangas
e474790400 Print more details on errors to log
Fixes https://github.com/zenithdb/zenith/issues/661
2021-10-01 17:57:41 +03:00
Alexey Kondratov
2c99e2461a Allow usage of the compute hostname in the proxy 2021-10-01 16:24:35 +03:00
Stas Kelvich
cf8e27a554 Proxy: pass database name in console too 2021-10-01 14:27:52 +03:00
Kirill Bulatov
287ea2e5e3 Limit concurrent relish storage sync operations 2021-10-01 08:37:09 +03:00
Heikki Linnakangas
86e14f2f1a Bump vendor/postgres 2021-09-30 20:36:57 +03:00
Arseny Sher
adbae62281 Rename SharedState.commit_lsn to notified_commit_lsn.
ref #682
2021-09-30 17:29:15 +03:00
Egor Suvorov
3127a4a13b Safekeeper::Storage::write_wal: clarify behavior (#679)
It previously took &SafeKeeperState similar to persist(), but only for its
`server` member.
Now it takes &ServerInfo only, so there it's clear the state is not persisted.
Also added a comment about sync.
2021-09-29 19:58:30 +03:00
Egor Suvorov
6d993410c9 docs/README: fix link to walkeeper's README (#677) 2021-09-29 14:40:16 +03:00
Kirill Bulatov
fb05e4cb0b Show better error messages on pageserver failures 2021-09-29 01:55:41 +03:00
Egor Suvorov
b0a7234759 pageserver: fix stale default listen addrs
* In command line help
* In dummy_conf
2021-09-28 20:57:51 +03:00
Egor Suvorov
ddf4b15ebc pageserver: use const_format crate to generate default listen addrs 2021-09-28 20:57:51 +03:00
Egor Suvorov
3065532f15 pageserver: fix mistype in listen-http arg help 2021-09-28 20:57:51 +03:00
Arthur Petukhovsky
d6fc74a412 Various fixes for test_sync_safekeepers (#668)
* Send ProposerGreeting manually in tests

* Move test_sync_safekeepers to test_wal_acceptor.py

* Capture test_sync_safekeepers output

* Add comment for handle_json_ctrl

* Save captured output in CI
2021-09-28 19:25:05 +03:00
Arseny Sher
7a370394a7 Wait till previous victim recovers in run_restarts_under_load.
Fixes test flakiness, as recovery easily might take the whole iteration.
2021-09-28 19:15:41 +03:00
Stas Kelvich
0f3cf8ac94 Cleanup Dockerfile.
* make .dockerignore `ncdu -X` compatible to easily inspect build context
* remove cargo-chef as it was introducing more problems than it was solving
* remove rocksdb packages
* add ca-certs in the resulting image. We need that to be able to make https
  connections from container with proxy to the console.
2021-09-28 18:26:20 +03:00
Heikki Linnakangas
014be8b230 Use Iterator, to avoid making one copy of page_versions BTreeMap
Reduces the CPU time spent in checkpointing, in the write_to_disk()
function.
2021-09-27 19:28:02 +03:00
Heikki Linnakangas
08978458be Refactor write_to_disk, handling dropped segment as a special case.
Similar to what commit 7fb7f67b did to 'freeze', dealing with the
dropped segment separately from the rest of the logic makes the code
easier to follow. It is also needed by the next commit that replaces
the code to build new BTreeMap with an iterator; we cannot pass one
of two kinds of closures as argument, it has to always be the same one.
Having separate DeltaLayer::create() calls for the case of dropped
segment and the other cases works around that.
2021-09-27 19:23:32 +03:00
Heikki Linnakangas
2252d9faa8 Switch to RwLock in InMemoryLayer
Allows more parallelism basically for free.
2021-09-27 19:15:40 +03:00
Arthur Petukhovsky
22e15844ae Fix clippy errors (#673) 2021-09-27 18:59:30 +03:00
Konstantin Knizhnik
ca9af37478 Do not write WAL at pageserver (#645)
* Do not write WAL at pageserver

* Remove import_timeline_wal function
2021-09-27 14:15:55 +03:00
Stas Kelvich
aae41e8661 Proxy pass for existing users.
Ask console to check per-cluster auth info.
2021-09-27 11:56:43 +03:00