klsince commented on PR #14188: URL: https://github.com/apache/pinot/pull/14188#issuecomment-2400795218
@tibrewalpratik17 thanks for the link, but I don't think "Scenario 2: Segment ZK Metadata CRC != Segment CRC deepstore" was the issue said in this PR. In this case, when RT table have multiple consuming segment replicas, all of them can seal the segment locally and try to commit it, but only the fastest replica wins and it updates both ZK metadata and deepstore (so crc in ZK and deepstore are same). But because the text index file is not deterministic, the other slower replicas can have different crc values so they have to download segments from deepstore to replace their local ones. -- 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]
