Skip to content

Conversation

@tbg
Copy link
Member

@tbg tbg commented Jan 6, 2025

Checked out sysbench_test.go from master at 363af5b.

Once included in a patch release, this will allow BenchmarkSysbench to show up
in the microbench dashboard1.

Footnotes

  1. https://microbench.testeng.crdb.io/dashboard/?package=pkg%2Fsql%2Ftests&benchmark=Sysbench&unit=&repository=cockroach&branch=master&days=90&end=2025-01-06T16%3A07

rickystewart and others added 30 commits December 3, 2024 14:06
…-24.3-135329

release-24.3: dev: use `nolintonbuild` for cross-builds
…-24.3-134980

release-24.3: scbuild: run test on two cores
…-24.3-136463

release-24.3: sql/schemachanger: Exclude schema change wait time from idle session timeout
Changes:

 * [`a7d74870`](cockroachdb/pebble@a7d74870) wal: prevent negative durations

Release note: none.
Epic: none.
Fix #136317
For the v24.3 version, this test fails consistently since the version is
missing improvements to the disk accounting in AC, where we adjust for
disk error. Since we can't backport the fix, making this test manual
only on this branch will reduce the noise.

Fixes: #136559.

Release note: None
This commit adds the cancellation check to the top K sorter and the
merge joiner to be performed on every input batch they read. I think that
these two are the only buffering operators that currently don't have any
cancellation checks:
- for hash joins and hash aggregation we have the check when performing
the hashing
- for ordered aggregator, buffered window functions, cross join, external
sort and other disk-backed operators we do the check on each input batch
- for general sort we do the check in the PDQ sort of each column.

Release note (bug fix): CockroachDB now better respects
`statement_timeout` limit on queries involving the top K sort and merge
join operations.
Previously the kafka v2 sink did not support multiple seed brokers,
where the v1 sink did. This PR adds this support.

Fixes: #136616

Release note (general change): add support for multiple seed brokers
in the new kafka sink.
…-test

roachtest: make disk bandwidth test manual only
Update pkg/testutils/release/cockroach_releases.yaml with recent values.

Epic: None
Release note: None
Release justification: test-only updates
…-24.3.1-rc-136219

release-24.3.1-rc: diagnostics: ignore timeout errors; extend client timeout
…-update-release-24.3

release-24.3: Update pkg/testutils/release/cockroach_releases.yaml
Previously, the arrayVariadicBuiltin was not correctly set on the
DistsqlBlocklist, since the range loop was modifying a copy of the
overload. This patch sets correctly sets the DistsqlBlocklist
flag on this builtin.

Fixes: #136601

Release note: None
Previously, the active record test had a test flake on
CockroachDB::FixturesTest#test_create_fixtures. This patch adds that
test on the ignore list.

Fixes: #136460

Release note: None
Previously, TestBenchmarkExpectation would flake on the
AlterRegions/alter_empty_database_drop_region depending on if the
descriptors were cached. To address this, this patch modifies the range
to avoid flakes on cache misses.

Fixes: #136392

Release note: None
release-24.3: sql: avoid unnecessary singleflight requests for role membership cache
…-24.3-136437

release-24.3: logictest: remove unnecessary flaky assertion from synthetic_privileges test
…-24.3-136219

release-24.3: diagnostics: ignore timeout errors; extend client timeout
…-24.3-136647

release-24.3: sql: avoid data race in TestReacquireLeaseOnRestart
Epic: None
Release note: None
Previously if the workload failed in the c2c roachtests which then caused the
latency verifier to fail, only a latency error would get returned to the
client. This patch adds yet another channel to the roachtest driver to properly
surface the workload error.

A more elegant solution would have involved ordering Monitor.Wait() calls, but
moving those around would likely cause fallout.

Informs: #136091

Release note: none
This commit extends the change in bee163a to not remember that a replica was
the leader if it is not fortified when stepping down. This gives us time to
stabilize the following state as the v24.3 release is rolled out:

```
r.state = StateFollower && r.lead = r.id
```

Once this state is stabilized (within and above `pkg/raft`), we can remove this
special case.

Epic: None
Release note: None
…7d74870cb20

release-24.3: go.mod: bump Pebble to a7d74870cb20
Epic: none
Release note (enterprise change): Adds AWS_USE_PATH_STYLE param to S3 URI
parsing.
TestParallelIOMetrics is very flaky and relies on
altering timings, so skip it under duress.

Fixes #136397
Fixes #136352
Fixes #135681
Fixes #134937
Fixes #134188
Fixes #133335
Fixes #129917
Fixes #125295

Release note: None
…-24.3-136265

release-24.3: changefeedccl: change cdc latency hist bucket sizes
release-24.3: crosscluster/logical: type check dest tenant spec in planHookFn
…nt-side

Add a metric
(`distsender.rangefeed.catchup_ranges_waiting_client_side`)
that counts how many rangefeeds are waiting on the
client-side limiter to start performing catchup
scans.

Fixes: #134681

Release note (general change): Add a new metric
(`distsender.rangefeed.catchup_ranges_waiting_client_side`)
that counts how many rangefeeds are waiting on the
client-side limiter to start performing catchup
scans.
…-24.3-136608

release-24.3: dev: remove stale reference to `cockroach-oss`
celiala and others added 25 commits December 23, 2024 14:03
…-v24.3.2-137612

staging-v24.3.2: released CockroachDB version 24.3.1. Next version: 24.3.2
release-24.3: sql/types, sql/stats: fix SQLStringFullyQualified for arrays of UDTs
…-24.3-137744

release-24.3: sql: allow CLOSE CURSOR in read-only txns
…-24.3-137870

release-24.3: roachtest: mark more activerecord tests as flaky
Release note: None
Epic: None
Release justification: non-production (release infra) change.
…4.3.2. Next version: 24.3.3

Release note: None
Epic: None
Release justification: non-production (release infra) change.
…lease-24.3-v24.3.2-au7w

release-24.3: released CockroachDB version 24.3.2. Next version: 24.3.3
…3.2-to-release-24.3-glnw

merge staging-v24.3.2 to release-24.3: released CockroachDB version 24.3.2. Next version: 24.3.3
Update pkg/testutils/release/cockroach_releases.yaml with recent values.

Epic: None
Release note: None
Release justification: test-only updates
As explained in #121708, jepsen artifact collection requires bzip2,
which is not installed on Ubuntu 22.04. This was preventing
investigation of tests failures.

This PR introduces bzip2 as a new software that can be installed by
roachprod install, and installs it while setting up the cluster to run
jepsen.

Fixes: #121708

Epic: none
Release note: None
…d index

Previously the `SHOW CREATE TABLE` statement would
lose index information. The fix now allows `SHOW CREATE TABLE`
to show the correct information that can be repeatedly entered back
into crdb to recreate the same table.

Fixes: #136410
Release note (bug fix): Previously `SHOW CREATE TABLE` was showing
incorrect data with regards to inverted indexes.  It now shows the
correct data that can be repeatedly entered back into crdb to recreate
the same table.
This commit should decrease the probability of flakes in the
`hash_sharded_index` logic tests due to transaction retries by making
a transaction high-priority.

Fixes #137674

Release note: None
`ALTER BACKUP SCHEDULE` being run on a backup schedule whose URI
contains spaces would fail with error `ERROR: parse
"'gs://cockroachdb-backup-testing/kevin spacey?AUTH=implicit'": first
path segment in URL cannot contain colon`.

This was caused by the fact that despite the `FmtBareStrings` flag being
used, the space character is considered one of the "special characters"
outlined in its doc that causes the string to remain wrapped with
quotes. Passing this quoted string to `url.Parse` fails.

Fixes: #134861

Release note (bug fix): `ALTER BACKUP SCHEDULE` no longer fails on
schedules whose collection URI contains a space.
release-24.3: backup: fix alter backup schedule failing on uri with spaces
This reverts commit a5c782f.

Epic: CRDB-37584
Informs #133189.
Release note: Introduces the `storage.columnar_blocks.enabled` cluster setting
allowing opting into the columnar block sstable format.
…-24.3-138044

release-24.3: roachtest: install bzip2 for jepsen tests
…-24.3-134853

release-24.3: crosscluster/physical: avoiding copying SSTWriter
This reverts commit 93e51ba.

Epic: CRDB-37584
Informs #133189.
Release note: none
release-24.3: add columnar block cluster setting
release-24.3: server,metrics: stop recording GaugeVec and CounterVec to tsdb
…-24.3-138059

release-24.3: logictest: fix flake in hash_sharded_index logic test
…-24.3-138043

release-24.3: schemachanger: Fixed bug incorrect InvertedColumnKinds in the inverte…
The raft status can be nil if internalRaftGroup is nil. This is possible
in the middle of replica removal, after
disconnectReplicationRaftMuLocked releases Replica.mu.

Epic: none
Release note: none
…-24.3-138287

release-24.3: kvserver: fix nil deref in replica metrics
@blathers-crl
Copy link

blathers-crl bot commented Jan 6, 2025

Thanks for opening a backport.

Please check the backport criteria before merging:

  • Backports should only be created for serious
    issues
    or test-only changes.
  • Backports should not break backwards-compatibility.
  • Backports should change as little code as possible.
  • Backports should not change on-disk formats or node communication protocols.
  • Backports should not add new functionality (except as defined
    here).
  • Backports must not add, edit, or otherwise modify cluster versions; or add version gates.
  • All backports must be reviewed by the owning areas TL. For more information as to how that review should be conducted, please consult the backport
    policy
    .
If your backport adds new functionality, please ensure that the following additional criteria are satisfied:
  • There is a high priority need for the functionality that cannot wait until the next release and is difficult to address in another way.
  • The new functionality is additive-only and only runs for clusters which have specifically “opted in” to it (e.g. by a cluster setting).
  • New code is protected by a conditional check that is trivial to verify and ensures that it only runs for opt-in clusters. State changes must be further protected such that nodes running old binaries will not be negatively impacted by the new state (with a mixed version test added).
  • The PM and TL on the team that owns the changed code have signed off that the change obeys the above rules.
  • Your backport must be accompanied by a post to the appropriate Slack
    channel (#db-backports-point-releases or #db-backports-XX-X-release) for awareness and discussion.

Also, please add a brief release justification to the body of your PR to justify this
backport.

@blathers-crl blathers-crl bot added the backport Label PR's that are backports to older release branches label Jan 6, 2025
@cockroach-teamcity
Copy link
Member

This change is Reviewable

@blathers-crl
Copy link

blathers-crl bot commented Jan 6, 2025

Your pull request contains more than 1000 changes. It is strongly encouraged to split big PRs into smaller chunks.

🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is dev-inf.

@tbg tbg closed this Jan 6, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

backport Label PR's that are backports to older release branches

Projects

None yet

Development

Successfully merging this pull request may close these issues.