[GitHub] [kafka] hudeqi commented on pull request #13421: KAFKA-14824:ReplicaAlterLogDirsThread may cause serious disk usage in…
hudeqi commented on PR #13421: URL: https://github.com/apache/kafka/pull/13421#issuecomment-1478842301 Hello, for "potential exceptions", I did an experiment to simulate a disk failure, which eventually lead to the unexpected disk growing. For details, please refer to the corresponding comment in [jira](https://issues.apache.org/jira/browse/KAFKA-14824) . -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [kafka] hudeqi commented on pull request #13421: KAFKA-14824:ReplicaAlterLogDirsThread may cause serious disk usage in…
hudeqi commented on PR #13421: URL: https://github.com/apache/kafka/pull/13421#issuecomment-1477368341 > > potential exceptions here (although I haven't found it yet, it may be potential) > > Could you share more information with me? The unexpected disk growing is a critical bug so it would be nice to dig in it deeply. The bug you fixed last time will lead to unexpected infinite disk growth. I understand that as long as the partition is marked as failed, this will lead to this result, so I think it is a necessary defense measure. As for the potential errors, I only found the "offset mismatch error" you fixed so far. :) -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [kafka] hudeqi commented on pull request #13421: KAFKA-14824:ReplicaAlterLogDirsThread may cause serious disk usage in…
hudeqi commented on PR #13421: URL: https://github.com/apache/kafka/pull/13421#issuecomment-1477340172 > Why we need this PR if [KAFKA-9087](https://issues.apache.org/jira/browse/KAFKA-9087) had fixed the bug ( you mentioned in the jira)? Is there another potential bug? Or the bug fixed by [KAFKA-9087](https://issues.apache.org/jira/browse/KAFKA-9087) is not root cause? KAFKA-9087 solves the root cause of stopping fetch due to an "Offset mismatch" error thrown during "processPartitionData", but I thought about it, and there may be other potential exceptions here (although I haven't found it yet, it may be potential), It will also lead to the final result: the fetch stops and the log is not cleaned up, and finally the disk usage grows infinitely. To be more precise, this pr is a defensive measure, as I understand it. -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [kafka] hudeqi commented on pull request #13421: KAFKA-14824:ReplicaAlterLogDirsThread may cause serious disk usage in…
hudeqi commented on PR #13421: URL: https://github.com/apache/kafka/pull/13421#issuecomment-1476492254 hello, maybe you are interested in this issue? @chia7712 -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org