Believe it or not I started to experience this last week. Has happened to me twice now. 6.0 mp3. My media server runs fine but for what ever reason my Master/Media just quits talking to it's library. That and I got two status 84 errors last night. I have not seen one of those in months.

The Great Pumkin is coming.

Whoa, that's weird too!
Forgot to mention that I'm on 5.1 MP5 (on AIX 5.3).
Wonder what would cause the same type of thing on different versions ...

Solar flares maybe?

Further observation:
While shutting down the daemons, I noticed a lot of tldd and tldcd processes were still running. Seems unusual.

I am on Sol 9 (My favorite OS so far) I have some bplib, jobd, dblib and nbproxy services that do not want to shutdown. Oh and one nbsl service.

I will be calling support as soon as a couple of jobs finish up.

This is just getting weirder and weirder ...
Now ALL of my master/media servers are showing a solid AVR in the Device Manager (after a service/daemon restart).

The library looks perfectly good though.

Oddly enough I had something similar happen to one of my master servers. It stopped talking to the library. I guess they were on bad terms.
Anyway, the master servers hba card went bad and for some reason that caused the library itself to lost its device mappings. So after I changed the card the two couldn't communicate still, until I manually entered the device mappings in the library itself.

Now I'm not spooked easily, but this has got me going a little bit. Last Wednesday night everything flipped out here in our environment. It was like all communication between the media servers and the library just haulted. Everything failed because the start window had expired, but there were no obvious clues available. I brought the entire backup environment down and cleanly booted it all and we've been working just fine ever since. Does Veritas put some "cookies" into their code?!

"C is for Cookie and that's good enough for me!"

When strange things happen with our library, I call the Storagetek support guy and ask him, "OK, I give. What DID you do?"

Thats weird because I have just had the same problem with netbackup 6.0 showing AVR against two drives out of eight instead of TLD. Stopping and starting services has no effect. Backups attempt to write to the drive and attemots are made to mount media but just goes onto say requesting next new media. Can't run inventory on the robot device but works fine when accessing via web browser console !!!!

Well, I'll be darned.

Power cycled the library and everything eventually came back up.

Weirdest day yet ...