JukeBox Losses Inventory

[demola]

I have a backup server is a Sun E450 running Solaris 6. Networker V 6.1.1. with one ATL 7100 Tape Library (64 slots) with 4 DLT Tape drives

The legato server frequently looses inventory of tapes in the jukebox. When I load tapes and run an inventory, if I execute nsrjb or view the contents of the jukebox using the nwadmin interface I see all the tapes I expect to see in the box.

A few hours latter if I execute nsrjb again or view the contents of the jukebox through nwadmin interface I see only one or two tapes in the drive.
The result is that during backups the server is always requesting for writable taps whilst all required tapes are already loaded.

So far I have done the following:
Reseat the SCSI cables and terminators at the tape library and the server.
Deleted all the tape drives and the jukebox. Removed nsrjb.res and recreated the Tape Library and jukebox using jbconfig

The Library inventory was stable for 2 Days and then the problem started again.

Does anyone have a solution?

'Demola

 

[wallace88]

When the jukebox seem to lose its inventory, run the following from command line:

sjirdtag <b.t.l>

where <b.t.l> is the bus, target, and lun of the robotic arm, as reported from inquire. i.e, run inquire, look for a line that says autochanger, then record and use the scsidev@b.t.l.

The sjirdtag will show information regarding your tape drives, slots, robotic arm, and mailbox. In each case, it should show if a tape is in a particular area (look for med_pres=) and the bar code label for the tape.

Compare the outputs of nsrjb and sjirdtag. If they match, then there is likely a problem with the jukebox, otherwise it is likely a problem with NetWorker.