anuragp010 commented on PR #11716: URL: https://github.com/apache/ozone/pull/11716#issuecomment-6910907160
Thanks for the review and detailed pointers, @devmadhuu . I agree this scenario needs an acceptance test with a real older client, but can we take it up as part of a follow-up JIRA [(HDDS-17828)](https://issues.apache.org/jira/browse/HDDS-17828) rather than in this PR? The existing xcompat tests run one command to completion in one container. This scenario needs an old-client write that stops after the first block is allocated. While it is stopped, a new client changes the bucket policy, and then the old client resumes. That needs: - an old-client writer that can pause and resume, for example reading from a FIFO (named pipe), with a smaller ozone.scm.block.size so the second block triggers allocateBlock. - a signal between the old-client and new-client containers, so the policy change happens at a fixed point. - changes to lib.sh and possibly a separate cluster configuration which has multiple volumes with different storage types. That is a larger change than this fix. -- 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]
