[GitHub] [accumulo] cshannon commented on issue #3023: External Compaction cancellation due to tablet split is very verbose

[GitBox <gi...@apache.org>]

cshannon commented on issue #3023:
URL: https://github.com/apache/accumulo/issues/3023#issuecomment-1304555822

   I started looking into this and there are 3 places where the stack traces are showing up, two of them in `FileCompactor` when `FileSKVWriter` is closed and 1 in `Compactor`. 
   
   In `Compactor`, I think we can definitely handle that by just catching the `CompactionCanceledException` and logging the message as a debug instead of error and full stack trace. We could log it as info but there's already info messages indicating the compaction is goign to be cancelled before cancel is called on the JOB_HOLDER. This also is in line with the existing behavior of CompactableImpl.  https://github.com/apache/accumulo/blob/d5a35f2f59bf6c35369f1032189812f98021e3d2/server/tserver/src/main/java/org/apache/accumulo/tserver/tablet/CompactableImpl.java#L1292-L1293
   
   For the other two cases where errors are logged it's a bit tricky. There's no guarantee that the errors causing the file writer to fail closing are related to compaction being cancelled and could be another serious issue. I do think that logging the entire stack trace is a bit verbose either way and at the very least maybe only log the error message and keep the stack trace at debug level.
   
   I will submit a PR shortly with the proposed logging changes and then can discuss if the logging level changes makes sense or not.


-- 
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: notifications-unsubscribe@accumulo.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org