https://github.com/ceph/ceph/blob/19428c64d26b4faec85822b08771a37159f8442f/src/mon/OSDMonitor.cc#L14668-L14679 So it appears we are trying to manipulate a bucket that no longer exists in the current crushmap and that is considered a fatal error. Looking at that it looks like we might be able to get a much better idea of what's happening by gathering the monitor log at debug_mon=20 and debug_paxos=20 when this occurs. Are you in a position to be able to attempt to reproduce this with the above log settings?
Missed the 6.1 z1 window. Retargeting to 6.1 z2.