Description of problem: When running a scale down operation the FQDN lookup for service agent names can be too strict. While agent names are generally assumed to be the FQDN of the host, and that the deployment engines [heat, mistral, ansible, puppet] will provide identical information as what is defined when the service was registered, this input could have an incorrect name, short or otherwise, due to a mirriod of unforeseen issues: 1. overridden in config 2. set in inventory 3. bad configuration on the host 4. bad facts etc... To make our process more resilient we changed the hostname lookup to use a fuzzy search. This fuzzy search will use the provided server name as an anchor and match against the available names. If a duplicate name is discovered (like in the instance where two hosts have the same FQDN or there is an improper setting in config), an error will be raised explaining the issue so that operators has the ability to confidently resolve the problem. Upstream change: https://review.opendev.org/695998
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-2020:0283