You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@fluss.apache.org by "Jackeyzhe (via GitHub)" <gi...@apache.org> on 2026/08/05 13:46:54 UTC
Re: [PR] [FLUSS-3483][coordinator] Reject lake snapshot commits for dropped tables [fluss]
Jackeyzhe commented on PR #3494:
URL: https://github.com/apache/fluss/pull/3494#issuecomment-5192569890
Hi @luoyuxia , thanks for the simplification and the rebase!
I've reviewed the changes and they look good:
1. **Inline check** — collecting `unavailableTableIds` upfront and checking inline in the IO loop is cleaner than the separate helper. Agree with this approach.
2. **IT test only** — `testCommitLakeTableSnapshotV2RejectedAfterDropTable` covers the end-to-end RPC → coordinator → ZK flow well. Dropping the unit tests in favor of the integration test makes sense here.
3. One thing I noticed: `upsertLakeTable` still uses `creatingParentsIfNeeded()`, so the ZK-level defense against late-commit znode resurrection we discussed earlier is no longer in place. The coordinator-side check covers the common case (table already dropped before the commit event is processed), but the narrow window between collecting `unavailableTableIds` and the IO task execution remains. Should we track that as a follow-up issue, or is this considered acceptable for now given the partitioned-table complexity the ZK boundary fix introduced?
CI is all green. Happy to help if there's anything else needed before this can move forward.
--
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: issues-unsubscribe@fluss.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org