Description of problem: This bug is about upgrading later ceph packages post RH Ceph 2.0 version using rolling_update playbook in ceph-ansible. currently the issues are: rolling_update playbook should support systemd - this is being tracked upstream by https://github.com/ceph/ceph-ansible/issues/814 also, currently there is lack of information for the user to know where to mention the source and destination packages for the upgrade - it would be nice to have this information included in the rolling_update.yml. Version-Release number of selected component (if applicable): How reproducible: always Steps to Reproduce: 1. install ceph using ceph-ansible 2. modify rolling_update.yml to reflect the source and destination packages for the upgrade 3. ansible_playbook rolling_update.yml Actual results: error: TASK: [Gracefully stop the OSDs (Sysvinit)] *********************************** failed: [10.8.128.103] => {"failed": true} msg: Job for ceph.service failed because the control process exited with error code. See "systemctl status ceph.service" and "journalctl -xe" for details Expected results: OSDs should be restarted and ceph upgrade should be successful. Additional info:
Need Tamil to check leseb's fix upstream
retesting leseb's fix. will update on the status soon.
gmeno to provide link to patches
https://github.com/ceph/ceph-ansible/compare/master...rolling-update?expand=1
I just merged the github branch as Tamil reported that she was able to successfully perform an upgrade. Tamil anymore comments?
sure Sebastien, I was able to upgrade from 10.2.1 to 10.2.2 for testing purposed to make sure rolling_update playbook worked and It DID ! :) please let me know if we need to be doing anything else.
Tested ceph upgrades using from RH ceph v10.2.2-12 to RH Ceph v 10.2.2-19 using ceph-ansible rolling-update playbook and it worked fine! source version - [http://download.eng.bos.redhat.com/rcm-guest/ceph-drops/auto/ceph-2-rhel-7-compose/Ceph-2-RHEL-7-20160630.t.0/] destination version - [http://download.eng.bos.redhat.com/rcm-guest/ceph-drops/auto/ceph-2-rhel-7-compose/Ceph-2.0-RHEL-7-20160712.t.0/] Test setup : 3 mons and 9 osds on a 4 nodes setup magna011: mon magna066: 3 osds magna094 - 1 mon, 3 osds magna095 - 1 mon, 3 osds ceph-ansible version: ceph-ansible-1.0.5-25.el7scon steps: 1. install ceph v 10.2.2-12 by copying repos locally running ceph-ansible site.yml and set up a cluster. 2. set ceph_origin: distro and ceph_stable_rh_storage: true in /usr/share/ceph-ansible/group_vars/all 3. copy the repos [destination repos to be upgraded to] to /etc/yum.repos.d using yum-config-manager. 4. set serial hosts for osds and mons. in my case, it was 3 mon hosts and 3 osd hosts. 5. run ceph-ansible rolling_update playbook. make sure the ceph version is upgraded to desired version.
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/RHEA-2016:1754