We had to clear the queue due to an emergency today. We had noticed that a filesystem was extremely slow around 2 days ago, and had updated our Galaxy configuration to work around this. Unfortunately we failed to restart the handlers so new datasets continued to be written to that path, until the point at which it became an emergency and the system became completely unresponsive.
We cleared the queue in order to allow us to restart the handlers, unfortunately we did not have a way to re-schedule the jobs we had to kill.
We're sorry for any inconvenience this has caused, this should be a one-time issue. Please restart your jobs at your convenience.