sodonnel commented on PR #7009: URL: https://github.com/apache/ozone/pull/7009#issuecomment-2275563703
In your example, what do you do with EC-3-2, if 2 replicas are missing and one of the remaining 3 has a missing chunk? You cannot recover it. There is certainly a problem here. If things are working as designed from the client size, and it is abandoning these blocks due to a failed ACK from one of the replicas, then the block length at OM should be the smaller size. Then the problem is actually that we have truncated the block length to remove the duplicated chunks, but we have not truncated the checksums to match that, and that results in the failed recovery. For some of the failing blocks, can you check the size of the block on OM to see if it matches the smaller or larger size of the block across the replicas? -- 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]
