Zoltan Borok-Nagy has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/22980 )

Change subject: IMPALA-14075: Parallelize ExpireSnapshots.executeDeleteWith()
......................................................................


Patch Set 2:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/22980/2//COMMIT_MSG
Commit Message:

http://gerrit.cloudera.org:8080/#/c/22980/2//COMMIT_MSG@14
PS2, Line 14: org.apache.iceberg.util.ThreadPools.getWorkerPool()
The class ThreadPools is in the "core" package of Iceberg which means we should 
use it cautiously, as even minor Iceberg versions can introduce breaking 
changes to these classes/methods. The Iceberg project follows semantic 
versioning rules only for the iceberg-api package: "From 1.0.0 forward, the 
project will follow semver in the public API module, iceberg-api." 
https://iceberg.apache.org/releases/#110-release

That said, the functionality we rely on is quite simple, so it should be easy 
to rewrite if Iceberg breaks it.

I have more concerns about the usage of getWorkerPool(), which is also used 
during table loading: "The size of the worker pool limits the number of tasks 
concurrently reading manifests in the base table implementation across all 
concurrent planning operations."

This means a long-running EXPIRE SNAPSHOTS can make Iceberg table loadings 
slower which can also cause higher lags in the event processor.

I'd suggest using a new ThreadPool dedicated for the EXPIRE SNAPSHOTS 
statement, or one that is shared between all EXPIRE SNAPSHOTS statement.



--
To view, visit http://gerrit.cloudera.org:8080/22980
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I6dcbf1e406e1732ef8829eb0cd627d932291d485
Gerrit-Change-Number: 22980
Gerrit-PatchSet: 2
Gerrit-Owner: Riza Suminto <[email protected]>
Gerrit-Reviewer: Daniel Becker <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Noemi Pap-Takacs <[email protected]>
Gerrit-Reviewer: Riza Suminto <[email protected]>
Gerrit-Reviewer: Zoltan Borok-Nagy <[email protected]>
Gerrit-Comment-Date: Thu, 05 Jun 2025 14:22:55 +0000
Gerrit-HasComments: Yes

Reply via email to