I've got a problem with a newly setup media server but it's a long tale so
make yourself comfortable before reading on !

Our company has been doing a network migration which has now been completed.
We used to have 2 backup servers (Sun E450s) with one as a Master/Media
server with a Storedge L9 attached and the other as just a Media server with
another L9.
One got migrated early to the new network so I set them both up as Masters
and ran them independently for a month.

The last backup server to migrate is leased and due to expire soon, so I
installed
another machine as a media server and added the L9 library to it.
During the install I said it wasn't a master server and supplied the name of
the master. That all went fine and the bp.conf and vm.conf on the master are
updated and reflect the name of the new media server.
I ran vmconf to setup the library and that went okay too.
I can manipulate the library with robtest, and I can use tar to write and
read to tapes.
I used bpadm on the new media server to define a storage unit, and that now
appears on the master. I also used device management on the master to tell
it to monitor the media server and that seems okay too.
If I run bpstulist on either server, it shows the original and new storage
units.
I loaded tapes into the "new" L9 and inventoried them from the master
server. They got correctly allocated to pools in accordance with the barcode
rules on the master.

I setup a test backup class to backup the new media server. I set it to use
"any available storage unit" and when I submitted it, it mounted a tape in
the L9
on the master server and backed up fine. I then configured the class to use
the new storage unit and submitted the job again.
It sits in the activity monitor with a status of "queuing" and never does
anything. It doesn't fail and it certainly doesn't go active.
In the activity monitor display, there is no entry in the media server
column for
this job, but then it hasn't gone "active" so maybe that's normal.

I created the log directories on both servers and get reams of messages, but
the only ones that stand out are 4 related to the job in question (1341)
[1341] <4> connected_peer: Connection from host escux04, 10.233.105.96, on
non-reserved port 42394
[1341] <4> process_request: VSMInit() failed for stunit change: 2d
[1341] <2> stunit_db: Q_STUNITGET
[1341] <4> bpdbm: request complete: exit status 0


Any advice or suggestions would be gratefully received. It's not the end of
the world if I can't sort it out. Licences etc. for 4.5 are on their way so
the current setup only has to function for another month or so. It just
riles me that I can't find what it is I've done wrong or overlooked.
If all else fails I'll run them as independent master servers again, but I'd
prefer not to.

Thanks for reading this far, and for all the useful hints I pick up from
reading the list traffic.
Regards,  Fergus