I forget if this had a dedicated thread.
Often, I have a situation where something (cloning, updating, etc) is blocked because backups are running — but the backups on a large server can take many hours, in our case half a day or more.
Is there a way that the backup system can only block actions on the actual app being backed up, and not all actions for all apps?
Having this constraint causes some pretty fundamental system design reconfiguration considerations, that could all be avoided if this little blocker to parallel processing of tasks were possible.
Yes, while I was demoing an app to someone on the demo server, someone managed to send it into a tarpit where all tasks were stuck pending a backup/cleanup task that would never complete.
I had to stop the backup/cleanup task and then wait for the 20 or so de/installs/app-changes to catch up.
Saved the demo gods from facepalming for too long