Description of problem: When upgrading from OSP13 to OSP16.1 an InstanceHA compute node, any pcs command executed on the already upgraded environment fails with: Error:Unable to read /etc/corosync/corosync.conf: Is a directory When the /etc/corosync directory was checked we could see that /etc/corosync/corosync.conf was really a directory. This directory was present also on all the other computes which were not upgraded yet, however in OSP13 it seems this wasn't a problem for pacemaker. Therefore, we need to add an upgrade task that checks if the directory exists and remove it if it's empty or back it up if it isn't. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. Deploy OSP13 environmetn with instanceHA 2. Add directory /etc/corosync/corosync.conf: mkdir /etc/corosync/corosync.conf 3. Run the FFU from 13 to 16.1 for control plane 4. Upgrade first compute 5. Run pcs status Actual results: Cluster check "pcs status" command fails on the compute node Expected results: Cluster check "pcs status" command works on the compute node Additional info:
Bugzilla not marked as a blocker as it can be easily workaround. The fix is to remove the /etc/corosync/corosync.conf folder (or simply rename it) from the Compute node.
Code present in RHOS-16.1-RHEL-8-20210120.n.1, validated in pidone instanceHA FFU CI job: http://10.19.227.143:8080/view/CI%20Jobs/job/osp-13-16-ffu-iha/105/
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 (Red Hat OpenStack Platform 16.1.4 director bug fix 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-2021:0817