Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

kv: hand off a number of roachtests #82060

Closed
tbg opened this issue May 30, 2022 · 7 comments · Fixed by #86300
Closed

kv: hand off a number of roachtests #82060

tbg opened this issue May 30, 2022 · 7 comments · Fixed by #86300
Assignees
Labels
A-kv-replication Relating to Raft, consensus, and coordination. C-enhancement Solution expected to add code/behavior + preserve backward-compat (pg compat issues are exception)

Comments

@tbg
Copy link
Member

tbg commented May 30, 2022

At the time of writing, KV owns many roachtests:

bin/roachtest list tag:owner-kv | grep -cF '[kv]'
220

We expect that for a lot of them, we can transfer ownership at some point in the near future.

I prepared a spreadsheet with a v0 of proposed movement: https://docs.google.com/spreadsheets/d/1w_zx8W1ObHVL1oMXNjh4LTcEMkTHyCycbGSnji1BvCk/edit#gid=2080879335

image

cc @lidorcarmel (for visibility) @srosenberg (since TE is the would-be owner for a lot of the tests and we should discuss the specifics of such a handover). Note that PERF is just TE today, but I thought it would make sense to make that distinction early.

Jira issue: CRDB-16193

@tbg tbg added the C-enhancement Solution expected to add code/behavior + preserve backward-compat (pg compat issues are exception) label May 30, 2022
@blathers-crl blathers-crl bot added the T-kv KV Team label May 30, 2022
@nvanbenschoten
Copy link
Member

@tbg what are the next steps here? To determine which tests can be handed off? To relabel those tests?

@tbg
Copy link
Member Author

tbg commented Jun 22, 2022

Test-Eng will take the perf tests mid-July and we'll hand of another batch after that (details TBD), possibly the version-upgrade tests.

@mwang1026
Copy link

Removing KV then since seems like Test Eng will take the lead on actually executing the change. Let me know if this is wrong @srosenberg

@mwang1026 mwang1026 removed the T-kv KV Team label Jul 11, 2022
@tbg
Copy link
Member Author

tbg commented Jul 12, 2022

I was going to make the PR.

@blathers-crl
Copy link

blathers-crl bot commented Jul 12, 2022

cc @cockroachdb/replication

@blathers-crl blathers-crl bot added the A-kv-replication Relating to Raft, consensus, and coordination. label Jul 12, 2022
tbg added a commit to tbg/cockroach that referenced this issue Jul 12, 2022
tbg added a commit to tbg/cockroach that referenced this issue Jul 12, 2022
tbg added a commit to tbg/cockroach that referenced this issue Jul 12, 2022
tbg added a commit to tbg/cockroach that referenced this issue Jul 12, 2022
tbg added a commit to tbg/cockroach that referenced this issue Jul 12, 2022
tbg added a commit to tbg/cockroach that referenced this issue Jul 12, 2022
craig bot pushed a commit that referenced this issue Jul 18, 2022
84265: roachtest: hand tpcc tests to test-eng r=srosenberg a=tbg

Touches #82060.

Release note: None


Co-authored-by: Tobias Grieger <[email protected]>
craig bot pushed a commit that referenced this issue Jul 18, 2022
84262: roachtest: test-eng owns perf tests r=srosenberg a=tbg

As of this PR, the test-eng teams owns its first 60 roachtests. The KV team now
owns ~160, down from ~220 initially. These tests are fairly stable, when they flake
it is mostly due to cloud infrastructure issues.

Touches #82060.

- roachtest: own kv perf tests to test-eng
- roachtest: own sysbench tests to test-eng
- roachtest: own tpccbench to test-eng
- roachtest: own tpc-e to test-eng
- roachtest: own ycsb to test-eng


Co-authored-by: Tobias Grieger <[email protected]>
@tbg
Copy link
Member Author

tbg commented Aug 16, 2022

Still need to talk the schedule and selection through with @srosenberg but this is the rest that I believe should be owned by test-eng:

  • Jepsen - if they fail then mostly due to flakes (apt-get etc). For actual test failure tutorial can reach out to KV if it fails (the average KV person doesn’t know how to investigate these either, but Tobi or Nathan do)
  • tpcc/headroom/n4cpu16
  • tpcc/mixed-headroom/n5cpu16
  • tpcc/w=100/nodes=3/chaos=true
  • version/mixed/nodes=3
  • version/mixed/nodes=5
  • weekly/tpcc/headroom

tbg added a commit to tbg/cockroach that referenced this issue Aug 17, 2022
See cockroachdb#82060.

Release justification: testing-only change
Release note: None
@tbg
Copy link
Member Author

tbg commented Aug 17, 2022

Here's what's left after #86300

$ roachtest list | grep -F '[kv]' | awk '{print $1}' | grep -Eo '^[^/]+' | uniq -c | sort
   1 autoupgrade
   1 drain-and-decommission
   1 drop
   1 gossip
   1 hotspotsplits
   1 inconsistency
   1 indexes
   1 ledger
   1 network
   1 overload
   1 queue
   1 restart
   1 slow-drain
   2 copy
   2 copyfrom
   2 kv50
   2 replicagc-changed-peers
   2 roachmart
   3 replicate
   3 transfer-leases
   4 loqrecovery
   4 splits
   5 decommission
   5 rebalance
   6 clock
   6 kv
   8 kv0bench
  10 decommissionBench
  11 acceptance
  13 follower-reads

@tbg tbg self-assigned this Aug 17, 2022
@tbg tbg self-assigned this Aug 17, 2022
craig bot pushed a commit that referenced this issue Aug 17, 2022
85277: roachtest: add mixed version testing for CDC. r=srosenberg a=renatolabs

This commits adds the `cdc/mixed-versions` roachtest. It builds on top
of the existing infrastructure for mixed version testing in order to
create a test scenario that reproduces the issue observed by DoorDash
when upgrading from 21.2 to 22.1.0.

Following the pattern present in other mixed version tests, this test
starts the cluster at the previous version, upgrades the binaries to
the current version, rolls it back, and then finally performs the
binary upgrade again and allowing the upgrade to finalize.

This roachtest uses the `FingerprintValidator` infrastructure used in
CDC unit tests (and also in the currently skipped `cdc/bank`
roachtest). This validator gives us strong guarantees that the events
produced by a changefeed are correct even in the face of ongoing
upgrades.

This test fails roachtest on v22.1.0; once the upgrade to v22.1 is
finalized, the Kafka consumer won't see further updates, and the test
eventually times out.

Release justification: test-only changes.
Release note: None.

86126: go.mod: upgrade twpayne/go-geom r=rafiss a=otan

This fixes ARM specific bugs.

Release note (bug fix): Intersection spatial operations would previously
potentially return incorrect results on ARM. This is now resolved.

Release justification: bug fix for existing functionality

Resolves #72226

86233: sql: remove super regions version and gates r=postamar a=RichardJCai

Release justification: Only removing old version.
Release note: None

86278: sql: udf logic tests not to skip dropping databases r=chengxiong-ruan a=chengxiong-ruan

Now that we support drop functions in legacy schema changer
and declarative schema changer falls back to legacy when seeing
any function descriptor, we can move udf logic tests back
to normal configs.

Release note: None.
Release justification: test only change.

86300: roachtest: own a few tests to test-eng r=srosenberg a=tbg

See commits.

Closes #82060.

Release justification: testing

86318: scplan: make EXPLAIN target output stable r=postamar a=postamar

This commit does not make any functional changes. It merely ensures that
the diagrams generated by an EXPLAIN (DDL) statement list the schema
change targets in a predefined order: first the to-public targets, then
the transient targets, and finally the to-absent targets.

This helps readability and makes the declarative schema changer
data-driven test output more stable.

Release justification: low-impact, test-only change
Release note: None

86331: roachtest: enable better vmodules for drain/quit roachtests r=kvoli a=aayushshah15

This should help in debugging failures like #85203 when they happen again.

Release note: None

Release justification: testing only

Co-authored-by: Renato Costa <[email protected]>
Co-authored-by: Oliver Tan <[email protected]>
Co-authored-by: richardjcai <[email protected]>
Co-authored-by: Chengxiong Ruan <[email protected]>
Co-authored-by: Tobias Grieger <[email protected]>
Co-authored-by: Marius Posta <[email protected]>
Co-authored-by: Aayush Shah <[email protected]>
@craig craig bot closed this as completed in b1811c9 Aug 17, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
A-kv-replication Relating to Raft, consensus, and coordination. C-enhancement Solution expected to add code/behavior + preserve backward-compat (pg compat issues are exception)
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants