Description of problem: Job starvation can occur when :drb strategy is used for RefreshWorkers Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Start EVM 2. Add Azure or Amazon provider (any provider with consolidated workers) 3. Start queueing the provider refresh (over long period of time, in some interval shorter than the time it takes to refresh Cloud manager) Actual results: Refresh of dependent managers (like Network or Storage) never happens Expected results: Dependent managers should be refreshed in order of job arrival to MiqQueue Additional info: Required to be fixed by the same time as the Azure targeted refresh lands in downstream [1] [1] https://bugzilla.redhat.com/show_bug.cgi?id=1487602
Also affects other providers with targeted refresh like OpenStack and AWS. Fixed by PR: https://github.com/ManageIQ/manageiq/pull/17187
PR merged.
Verified on 5.10.0.18.