Description of problem: When new OVS version 2.5-23 and UP is consumed openvswitch is not restarted , but new OVS is installed. (https://bugzilla.redhat.com/show_bug.cgi?id=1403958) cloud admin should be notified to restart all components related to OVS or reboot the host. otherwise he will end up with new OVS installed but old OVS running. I would be best if OVS could send some notification to all neutron/other services to be restart or to find different solution. I Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Either upgrade or update, I think the overall OSP process needs some thinking because the issue isn't specific to OVS. For instance, let's say we have a security fix in glibc or openssl. How do you recommend customers to proceed? There is a plugin for yum to tell if restart is needed (needs-restarting). Although it is a much better than something specific to OVS, I suggest to consider even further, like a process to migrate all the workload out of the compute node, take it out of production, upgrade or update, reboot, get it back to production and re-balance the workload.
The Alert is required on undercloud update & overcloud update
@Assaf, any idea on how such a thing can be implemented? Can we just add a warning after the upgrade command? Thanks, Nir