MEDIA LOAD OR EJECT FAILED

I am not sure what you have - is this a standalone tape drive or a jukebox drive?

It could be that the tape cannot be rewound but it is also possible, that it has not even been loaded yet during a jukebox operation.

If it is a jukebox drive, I would try to reset the jbox first (nsrjb -HEv). Then retry the operation. If this does not work, recycle the power on the jukebox and the drive (in case they have different power supplies).

If this does not work, manually load a scratch tape in the drive and verify that you can read (use scanner for that purpose). If yes, try to label it. It this is successful, you most likely have a jukebox, not a drive problem.

You of course may also quiese NW and use standard OS tools to verify whether the drive behaves fine.

standard SCSI Jukebox

I just hope that you are sitting next to it and easily can do what I suggested.

does the tape library have its own web interface, showing you where a tape is actually located, to see if NW is out of sync?

When we were still using physical libraries (before switching to Data domain using them as virtual tape libraries, but nowadays we only use ddboost, so no fibre connectivity involved anymore more, only backup over lan or wan towards DD systems), I also used sji-tools heavily, like sjirdtag and sjimm to be able to perform activities outside of NW itself as having to reset drives or the whole jukebox (with at times hundreds of defined tape drives and possibly thousands of tapes), was way too cumbersome.

Also when testing a drive I'd use nsrjb -Lnv to mount-but-load a tape into a drive and then use nsrmm -pv -f to read the lable of the tape. When trying to mount it could get much easier stuck again, needing to go into timeouts and so on. Doing nsrjb and nsrmm after eachother was a very quick way to test drives if they were fine (also way back in the day easy to spot if device files would have changed, before NW itself got better at discovering and dealing with that).

You need to get to know how to determine if possibly a tape or a drive (or even the robotics arm) is no longer working by testing for example the same tape in a different drive, thus ruling out the tape is an issue, but rather the drive is?

Depending on how fancy the library is, it might even give you tools to look at and report about drive and tape health (or IBM libraries could do that, creating pivot tables in excel).

Networker itself is not the only interface towards the hardware. NW also provides various tools like the sji-tools, that can help to map out how things are doing. But any interface of the library itself also is very helpful as NW can only do so much, in case there is an issue on the library, either with a tape, drive or robotics arm...