Red Hat Bugzilla – Bug 1518666
RHCS upgrade 2.4 - 3.0: Fails with "msg": "file (/var/log/ganesha) is absent, cannot continue"
Last modified: 2018-02-22 07:27:04 EST
Created attachment 1360291 [details] upgrade log Description of problem: rolling upgrade fails while upgrading nfs client since ceph-ansible searches for /var/log/ganesha, but on the nfs client that particular file will be /var/log/ganesha.log Version-Release number of selected component (if applicable): RHEL - 12.2.1-40 How reproducible: Steps to Reproduce: 1. Configure 2.4 cluster with nfs(manually). 2. Upgrade to 3.0. Actual results: TASK [ceph-nfs : change ownership on /var/log/ganesha] ************************************************************************************************************************************************************ fatal: [magna067]: FAILED! => {"changed": false, "failed": true, "msg": "file (/var/log/ganesha) is absent, cannot continue", "path": "/var/log/ganesha", "state": "absent"} Additional info: I tried two workarounds: 1. Moved the /var/log/ganesha.log file to /var/log/ganesha on the nfs client. Reran the upgrade and it finished successfully. 2. Edited the path to "/var/log/ganesha.log" in "pre_requisite_non_container.yml" in ceph-nfs tasks. Reran the upgrade and it finished successfully. This bz has to be release noted.
Just to clarify, ceph-ansible is looking for a directory not a file. Usually /var/log/ganesha is the directory where ganesha puts its logs in. The right work around is to create the directory: /var/log/ganesha It is weird though, I suspect you are running an old version of ganesha. Bara, I changed the description.