Description of problem: In previous versions, these commands could non-deterministically crash osds if issued during active client writes. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. Produce unfound objects (see the ceph-qa-suite tests which do this) 2. Run client writes (like rados bench) 3. Run mark_unfound_[delete|revert] Actual results: Sometimes, an osd crashes. Expected results: The objects are either reverted or removed. Additional info:
Sam I think you said this issue is resolved upstream in v10.0.4? is that correct?
Yes
Thanks Sam!
See tasks/[ec_]lost_unfound.py and tasks/rep_lost_unfound_delete.py.
verified in 10.2.2-23.el7cp
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2016:1755
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 500 days