Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• Ensure there is a link to an issue (except for internal cleanup and flaky test fixes), new features should have an RFC that documents use cases and test cases.

Tests

• Bug fixes should have at least one unit or end-to-end test, enhancement and new features should have a sufficient number of tests.

Documentation

• Apply the release notes (needs details) label if users need to know about this change.
• New features should be documented.
• There should be some code comments as to why things are implemented the way they are.
• There should be a comment at the top of each new or modified test to explain what the test does.

New flags

• Is this flag really necessary?
• Flag names must be clear and intuitive, use dashes (-), and have a clear help text.

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow needs to be marked as required, the maintainer team must be notified.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from vitess-operator and arewefastyet, if used there.
• vtctl command output order should be stable and awk-able.

I got feedback to not move the flag to another package, but I get circular dependencies otherwise, not sure how to proceed here

I got feedback to not move the flag to another package, but I get circular dependencies otherwise, not sure how to proceed here

Sorry, meant to take a look at this earlier. My thought was to move the flag to tablet-level. I will take a look first thing tomorrow and push to this PR if that is OK.

I can take a look at the failing unit tests too then.

Sorry, meant to take a look at this earlier. My thought was to move the flag to tablet-level. I will take a look first thing tomorrow and push to this PR if that is OK.

This commit moves the flag to the vttablet level.

I can take a look at the failing unit tests too then.

Didn't look at the failing tests since I saw you had some recent commits to try to address them. Let me know if you want me to take a look at them.

Thanks for that commit, I've pulled it in! I took a stab at the tests, but I would love some help with the remaining ones

Thanks for that commit, I've pulled it in! I took a stab at the tests, but I would love some help with the remaining ones

Only one other test related to noblob seems to be failing. Here is the commit that fixes it: 99e1dfd. For noblob there is another set of queries that need the timeout hint, which I have added in that commit.

Can you pull that and let's see if all tests pass? Thanks.

Tests are all now passing! I'm also trying to find someone with permissions to update the hubspot repo settings so others can actually push to my branches 😅

@olyazavr, great, good to see that all tests are passing.

Following up on the slack thread regarding this issue: did this PR fix your existing issue? Especially the fact that we are adding the timeout hints on the target for the insert/updates. Usually we have been seeing issues on the source while selecting rows in the copy phase. So just wanted to confirm that it does work.

Can you also add your specific findings to the description in the issue, just so that we have it all in one place? Thanks!

Yes, it did fix the issue, and we are now able to run migrations on larger/busier keyspaceshards that previously could not run schema migrations on ghost or vitess (:

If the question is do we need the timeout hints for anything other than the selects - I confess I didn't try this out, and added the timeout hints in all places to be consistent, and because it seemed like there was no downside to having all vreplication queries be allowed to run longer

I did not test this with any other usage of vreplication other than online ddl schema migrations

Rebased and dealt with merge conflicts

Looks like all checks are passing, anything else for me to do here?

I will open a PR to update the docs here: https://vitess.io/docs/19.0/reference/vreplication/flags/#vreplication_copy_phase_duration

After this PR the flag is applicable on BOTH the source and target keyspaces:

1. On the target as it is used for the context used in the RPC to the source tablet by the vcopier:

   vitess/go/vt/vttablet/tabletmanager/vreplication/vcopier.go

   Lines 395 to 423 in 59408f9

   	ctx, cancel := context.WithTimeout(ctx, vttablet.CopyPhaseDuration)
   	defer cancel()
   	var lastpkpb *querypb.QueryResult
   	if lastpkqr := copyState[tableName]; lastpkqr != nil {
   	lastpkpb = sqltypes.ResultToProto3(lastpkqr)
   	}
   	rowsCopiedTicker := time.NewTicker(rowsCopiedUpdateInterval)
   	defer rowsCopiedTicker.Stop()
   	copyStateGCTicker := time.NewTicker(copyStateGCInterval)
   	defer copyStateGCTicker.Stop()
   	parallelism := getInsertParallelism()
   	copyWorkerFactory := vc.newCopyWorkerFactory(parallelism)
   	copyWorkQueue := vc.newCopyWorkQueue(parallelism, copyWorkerFactory)
   	defer copyWorkQueue.close()
   	// Allocate a result channel to collect results from tasks.
   	resultCh := make(chan *vcopierCopyTaskResult, parallelism*4)
   	defer close(resultCh)
   	var lastpk *querypb.Row
   	var pkfields []*querypb.Field
   	// Use this for task sequencing.
   	var prevCh <-chan *vcopierCopyTaskResult
   	serr := vc.vr.sourceVStreamer.VStreamRows(ctx, initialPlan.SendRule.Filter, lastpkpb, func(rows *binlogdatapb.VStreamRowsResponse) error {

2. On the source as it is used when building the query by the row streamer that runs there:

   vitess/go/vt/vttablet/tabletserver/vstreamer/rowstreamer.go

   Lines 458 to 460 in 59408f9

   	func GetVReplicationMaxExecutionTimeQueryHint() string {
   	return fmt.Sprintf("/*+ MAX_EXECUTION_TIME(%v) */ ", vttablet.CopyPhaseDuration.Milliseconds())
   	}

vitessio/website#1797