Hastyshell opened a new pull request, #61143:
URL: https://github.com/apache/doris/pull/61143
### What problem does this PR solve?
Problem Summary:
When a transaction is in TIMEOUT_SUCC state (partial publish success), the
partition's visible version advances while some base replicas may still lag
behind. Previously, schema change/rollup jobs would directly use the
partition's visible version as alterVersion without checking if the
corresponding base replicas have actually caught up locally. This causes BE to
fail with 'base tablet's max version < request version'.
This fix adds a gating check in the WAITING_TXN phase: before sending alter
tasks, verify that each base replica (corresponding to shadow replicas) has
caught up to the partition's visible version:
- replica.getLastFailedVersion() < 0
- replica.checkVersionCatchUp(visibleVersion, false)
If any base replica lags, the job stays in WAITING_TXN and retries in the
next cycle, allowing the lagging replica to catch up via normal publish or
repair mechanisms.
Changes:
- AlterJobV2: add isReplicaVersionComplete() helper
- SchemaChangeJobV2: add allBaseReplicasCaughtUp() gating before sending
tasks
- RollupJobV2: same gating for consistency
- Add unit tests for both scenarios
### Release note
None
### Check List (For Author)
- Test <!-- At least one of them must be included. -->
- [ ] Regression test
- [x] Unit Test
- [ ] Manual test (add detailed scripts or steps below)
- [ ] No need to test or manual test. Explain why:
- [ ] This is a refactor/code format and no logic has been changed.
- [ ] Previous test can cover this change.
- [ ] No code files have been changed.
- [ ] Other reason <!-- Add your reason? -->
- Behavior changed:
- [ ] No.
- [ ] Yes. <!-- Explain the behavior change -->
- Does this need documentation?
- [ ] No.
- [ ] Yes. <!-- Add document PR link here. eg:
https://github.com/apache/doris-website/pull/1214 -->
### Check List (For Reviewer who merge this PR)
- [ ] Confirm the release note
- [ ] Confirm test cases
- [ ] Confirm document
- [ ] Add branch pick label <!-- Add branch pick label that this PR should
merge into -->
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
