Hi Ed, Thanks for articulating some key points in the bug description and comments.
I had some questions about technical details and alternatives to fix this, but could find most of those answers in the linked bugs/commits/etc, and will summarize my findings here, in another comment. Summary is, I think there is no good way to "fix" this right now, and that is (as far as I could find) the same in openstack upstream, and fortunately the revert/change of default (to avoid the problem) is the least bad option, apparently, considering it's the value in prior _and later_ Ubuntu releases, and has been backported in openstack upstream to Yoga (shipped with Jammy). Question: the `Test Plan` should include a rabbitmq restart, in order to trigger the problem, as indicated in `Regression Potential`: "nova-compute services crash when rabbitmq-server is restarted" ? Request: could you please clarify the impact of the 'crash' (or actually a 'stall'?) in the `Impact` section? Because `crash` seems to indicate the program stops running, but per comment #26 it stays active (running). -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1934937 Title: [SRU] Heartbeat in pthreads in nova-wallaby crashes with greenlet error To manage notifications about this bug go to: https://bugs.launchpad.net/cloud-archive/+bug/1934937/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
