The most likely cause is the remote agent on one of the servers being backed up hanged. To troubleshoot, you may want to divide your backup job into multiple ones based on each server. When the job hangs next time, you know which server is involved and may try to restart that remote agent if possible. If not, you may have to reboot the server.
Yesterday evening I encountered the same problem. The job was already running for 18 hours while in normal situations this takes only 3 hours. I could not eject the tape. So I tried to cancel the job. Result: "cancel pending". This morning I went to check: still "cancel pending".
I've ejected the tape by pushing the eject button for a longer time. And I also tried to eject it by selecting eject media in the device screen. I don't know which one of these both actions led to the result, but it ejected anyway. So that problem is solved.
Now it looks like the job is still going on (now already 35 hours). How exactly can I delete the job that is running for the moment concerned that I've already canceled it?