luoyuxia commented on issue #1583: URL: https://github.com/apache/fluss/issues/1583#issuecomment-3392727448
> > [@luoyuxia](https://github.com/luoyuxia) thank you for the investigation! I think we can increase or remove the `log.replica.max-lag-time` configuration in the test, because we now supported controlled shutdown that the stopped tabletserver will send shutdown request to coordinator to trigger isr shrink at once without waiting the out-of-sync lag time. What do you think [@swuferhong](https://github.com/swuferhong) ? > > If we want to remove `log.replica.max-lag-time`, the test cases also need to be modified. This is because the current test does not trigger a `tabletServer shutdown` but instead triggers a `replica stop` to mock replica out of sync. However, I feel that modifying the test to trigger a `tabletServer shutdown` would no longer be meaningful for this test. +1, if shut down tabletserver, the server won't receive the `NotifyRemoteLogOffsetsRequest` and then won't delete the segments that already copied to remote. But the test want to test although a follower replica is out of sync, if it leader commit remote log segment, the follower replica should still delete the log segment. -- 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]
