I'm using dynamic workers allocation in our private openstack cluster through Openstack Cloud plugin 1.8.
I see Perforce Plugin cannot resolve host names of those new workers. I suspect it fails with NPE or something because it cannot checkout files to the new worker.
I have to delete the old workspace (leftover from another worker) and restart the build. this time it will work.
but if next build happens on a yet another new worker, I have to repeat this: delete the workspace and re-run the build.
so.. why not use IP address if hostname cannot be resolved? or just always use IP address for perforce workspaces. maybe transform it to something Perforce-friendly like 192-168-1-123.
this bug blocks us from using dynamic openstack workers with Perforce plugin