Description of problem: ceph-ansible fails when the osd handler is triggered. RUNNING HANDLER [ceph.ceph-common : copy osd restart script] ******************* fatal: [osd0]: FAILED! => { "failed": true } MSG: The conditional check 'item in play_hosts' failed. The error was: error while evaluating conditional (item in play_hosts): 'item' is undefined The error appears to have been in '/opt/guits/ceph-ansible/roles/ceph-common/handlers/main.yml': line 31, column 5, but may be elsewhere in the file depending on the exact syntax problem. The offending line appears to be: - block: - name: copy osd restart script ^ here Version-Release number of selected component (if applicable): How reproducible: It happens when a parameter that has been modified in group_vars/* results with a change in ceph.con so handlers are triggered. (Handlers are not triggered during the initial deployment). Steps to Reproduce: 1. Deploy a new cluster 2. Change a parameter in group_vars/osds (a parameter that makes a change in ceph.conf) 3. Launch ceph-ansible again. Actual results: The installer will fail when handlers are triggered. Expected results: The install doesn't fail, it updates the ceph configuration and restarts properly the ceph cluster.
- Fixed by https://github.com/ceph/ceph-ansible/commit/d8387cf82c647b85d51d5c28bd3a2c37f02ebd84 - Merged in master. - Backported in stable-2.2 - New tag v2.2.3
Vasishta, Would you please set qa_ack + so that we may attach the build containing the fix?
Unfortunately "v2.2.3" was tagged on the master branch, rather than the stable-2.2 branch. I suggest we delete the v2.2.3 tag upstream and recreate it on the stable branch at commit 2eb098ea5f0106f48f2dceca157ce6f5e4cda8b5. Guillaume and Andrew, are you ok with this plan? To move this bug forward I've pushed a new tag (v2.2.4) for the tip of stable-2.2 today and we'll ship that here.
(In reply to Ken Dreyer (Red Hat) from comment #7) > Unfortunately "v2.2.3" was tagged on the master branch, rather than the > stable-2.2 branch. > > I suggest we delete the v2.2.3 tag upstream and recreate it on the stable > branch at commit 2eb098ea5f0106f48f2dceca157ce6f5e4cda8b5. Guillaume and > Andrew, are you ok with this plan? > > To move this bug forward I've pushed a new tag (v2.2.4) for the tip of > stable-2.2 today and we'll ship that here. Sounds good to me. I've done this and the new v2.2.3 tag is pushed. https://github.com/ceph/ceph-ansible/tree/v2.2.3
Thanks!
Hi, Tried changing journal size on second run. It caused change in conf file and ceph-ansible didn't fail when the osd handler is triggered. RUNNING HANDLER [ceph.ceph-common : restart ceph osds daemon(s)] *************** changed: [magna068] ------------------------------ RUNNING HANDLER [ceph.ceph-common : copy osd restart script] ******************* changed: [magna071] changed: [magna106] RUNNING HANDLER [ceph.ceph-common : restart ceph osds daemon(s)] *************** changed: [magna071] changed: [magna106] Moving this BZ to VERIFIED state. Regards, Vasishta
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-2017:1496