smengcl commented on PR #6014: URL: https://github.com/apache/ozone/pull/6014#issuecomment-2161499786
> The failure in TestHsync is inherent in the test code. The PR just made it more likely to fail. > > The WriteChunk/PutBlock waits for ALL_COMMITTED. However, the fact that all peers have committed does not imply they are applied. In other words, the read after write is not linearizable. > > I see there are two solutions (1) enable Ratis linearizable read, and use XceiverClientRatis to read data (unsupported today) (2) Add a workaround at Ozone client side to wait and retry. Thanks @jojochuang for looking at the `TestHSync.testOfsHSync` failure. I wonder if it is the reduction in latency achieved in this PR (?) that amplified the issue in this case. cc @szetszwo -- 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]
