YenchangChan commented on issue #66299: URL: https://github.com/apache/doris/issues/66299#issuecomment-6551732522
Hi @furquan39, thanks for the detailed investigation and production evidence. I have encountered a similar repeated-compaction-failure problem in production. I had suspected that the fixed tablet_sched_delay_time_ms retry delay was involved, but your report provided a much more complete analysis of the failure and its impact. Based on your analysis and proposed direction, I put together an initial PR: [#69920](https://github.com/apache/doris/pull/69920). The PR tracks consecutive failures separately for cumulative and base compaction, applies a capped exponential backoff (5s, 10s, 20s, ... up to a configurable 5min by default), and resets the failure count after a successful compaction. The current scope is limited to retry backoff; it does not yet include the quarantine state or additional observability proposed in this issue. I would really appreciate your feedback, especially on the backoff policy and default cap. If you already have related implementation work in progress, I would also be happy to compare approaches and coordinate on the solution. -- 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]
