Networker Tape Drive Issue

This is really a puzzling one as at least one piece is missing. Is this a silo which is shared among multiple NW servers?

With respect to the win32 errors these are obviously a NW/Windows servers. The error you see is not a NW error but a Windows error. Simply run "net helpmsg 1" and you will receive the same message:

D:\>net helpmsg 1

Incorrect function.

D:\>

The key question is whether this is a drive or a jukebox related problem. You may want to load a test tape into the correct drive and then run an appropriate NW command to verify that.

  -  Run jbexercise to verify whether the silo will behave as expected.

  -  Run nsrjb to load (not mount) a labeled tape in a drive of your choice.

  -  Run the command scanner -n -m -vv device on the drive to verify if the tape can be read.

  -  Run nsrjb to label a test tape.

See the Command Line Reference for details.

If you cannot move the tapes as expected, this is obviously a NW error.

If you have read/write errors, this is obviously a tape drive related error (hardware and/or driver). As NW does not use an extra device driver, you will not even succeed with any other backup software you may want to try.

Please also see the daemon.raw file for more detailed error messages.

Instead of using scanner to read the whole tape, trying to only read the tape label when you loaded (and not mounted) the tape into a drive, might show if NW can actually read the mounted tape.

something akin to:
# nsrjb -lnv -f -S slot -j

and the use nsrmm to read the label only

Of course it does not make sense to read the whole tape.

  -  If there is a general read error, I expect the process to fail right away.

  -  If scanner will start successfully, you can safely abort the process. You media index will not be affected if you use the -n option as suggested.

Of course you could also use nsrmm to verify/(re-)label a tape. But be careful - as far as I remember, this only works for standalone devices and not for jukebox devices. Of course, I could be wrong ...

Some years ago for myself really, after we made the transition to datadomain, so no hassling with jukeboxes and (physical and virtual) tapedrives anymore.

We used this approach all the time, for example after tape library planned maintenance or after issues, with a script that would use nsrjb to load-no-mount a tape from a specific slot into each and every drive and then use nsrmm to only (try to read) the volume label of this tape that was labeled into a test-only pool, not used for anything else really. So no matter how many drives we had, the script would go through all of them.

With physical tape libraries we had drives shared, so you'd be testing some drives multiple times bit from different hosts (or used a script that only tests it in the first system) but later on with vtl's without drive sharing, you'd be testing each individual drive only once.

Using load-no-mount prevented having to reset a drive (-HH -v) as NW would think a tape was in the drive while there wasn't any (or even another tape if device files on storage node end might have been non-persistent and reshuffled again). Never understood why resetting a drive always had to take so long, especially when you were sure another tape or no tape was in the drive, which one can validate from library end (if it has a gui) or by using the NW sji tools like sjirdtag and sjimm and the like to perform mount commands and look at the library slots and drives what is loaded where... very useful as you are testing outside of networker itself but still from the nw server (or from a nw storage node controlling the jukebox robotics arm).

So I expect nsrmm to work to only read the label, a very basic functionality test only however, preventing having to reset a drive or jukebox when drive or tape is broken.

Ideally you would have set the drive in service mode, so it wouldn't be used by NW. However I can recall some nw versions used the service mode not in a way we'd have hoped for.

This we did not test on any recent nw versions, so behaviot/functionality might have changed. Mainly used up until nw7 and 8 and maybe a tiny bit in nw9.

reading the tape of a loaded-not-mounted tape in a drive would be (if memory serves me well):

# nsrmm -pv -f

Just hat the chance to get hold of a jukebox and to verify the labeling behavior.

Unfortunately, I was mistaken. NW will still be able to label a loaded tape with nsrjb although it is a jukebox tape drive. For verification I used NW 19.5.0.4