Description of problem: The check of PGs is not really efficient because the current check might be satisfied as long as 1 PG is active+clean. How reproducible: The check occurs when the handler are triggered after any modification has been brought to ceph.conf. Steps to Reproduce: 1. Deploy a new cluster 2. Make a change 3. run the installer again so the daemons are restarted (OSD in this case) Actual results: When the OSD daemons are restarted, the check will be satisfied as long as the pattern 'active+clean' is found in the ceph status output. Expected results: The check is satisfied only if the total number of PGs are in 'active+clean' state. fix: https://github.com/ceph/ceph-ansible/pull/1580 backport: https://github.com/ceph/ceph-ansible/pull/1581
We have this new approach new: https://github.com/ceph/ceph-ansible/pull/1584
The PRs are merged upstream to master, so this will be fixed in the upcoming ceph-ansible v2.3.0.
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-2017:3387