Comment 6 for bug 1784579

Unfortunately there are no previous errors in the logfile. It looks like something bad happened during the upgrade. Because we had some trouble with our rabbitmq and duplicate messages. So maybe the live migration request has been cached and after fixing the duplicate message problem (we had to reinitialize our rabbitmq cluster) the instance cache info was still wrong (state: migrating to). The only way to solve that errors above (NovaException_Remote: Unsupported VIF type ...) was to remove that wrongly cached port from the instance. With newly created instances there was no problem. What I don't understand is where does this information come from? Because I updated the wrongly cached information in the database with information before the upgrade. And as soon as I started a new action (e.g. reboot) the instance cache information had the state: migrating to. So where does this come from?