Public bug reported: This is to track and SRU the upstream bug [0]. In large clusers, there's a GIL deadlock that happens (or observable) when ceph-mgr receive high number of requests from Prometheus module.
This bug likely exists at least since the initial release of Nautilus. Upstream fixed this recently [2]. Targetted for Octopus 15.2.9 (which is not released yet - the next point release). So this tracker is to assess the impact and potentially backport in Ubuntu/UCA packages before 15.2.9 gets released. N.B: We have previously backported a similar bug [1]. But this is unrelated. [0] https://tracker.ceph.com/issues/39264 [1] https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1906496 [2] https://github.com/ceph/ceph/pull/38677 ** Affects: ceph (Ubuntu) Importance: Undecided Status: New ** Tags: seg ** Tags added: seg -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1917494 Title: ceph-mgr hangs in large clusters To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1917494/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs