graysonzeng commented on issue #21860: URL: https://github.com/apache/pulsar/issues/21860#issuecomment-1894889891
> @graysonzeng Thanks for reporting and providing details about the scenario. It seems that this bug is similar to #20526, #5284 and #14941 . I checked your PR 4171 in apache/bookkeeper, but couldn't find how it would change the behavior since all calls to sendAddSuccessCallbacks already happen in a synchronized manner. Would you mind taking a look at #20540 ? That addresses race conditions in Pulsar related to topic loading and unloading. I haven't yet found a consistent way to reproduce the issue. I now have an idea to do this rolling restart for bookkeepers based on your issue description. Reproducing the issue has been the main limiting factor by now so contributing a test case that reproduces the issue would be extremely helpful as the starting point. I'll check it out once I finish the other fixes -- 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]
