Veritas-bu

[Veritas-bu] Incorrect Tape Drive Status Causing Drives To Fail on

2006-12-04 23:56:12
Subject: [Veritas-bu] Incorrect Tape Drive Status Causing Drives To Fail on
From: ddunham at taos.com (Darren Dunham)
Date: Mon, 4 Dec 2006 20:56:12 -0800 (PST)
> I have a StorageTek L180 library with IBM LTO-2 drives connected to a
> Sun media/master server via fibre channel. I have six drives and I
> recently migrated to a new media/master server. I'm having a problem
> with two of the drives in the system, they show up at the PROM when I
> run 'probe-scsi-all' and in the OS when I run 'cfgadm -al' and 'sgscan
> tape' and show up in robtest. However when I load a tape into them
> with robtest, or when NetBackup loads a tape into them the tape
> seemingly disappears, a media not present error is returned and the
> drive downs itself.

That can happen if the drive mapping is inconsistent between the solaris
devices and the library ordering.

> I have rebuilt the drives starting at the fiber channel level by
> deleting the Fabric_WWN_Map file and running cfgadm with the
> 'force_configure' option, rebuilt my rmt drivers by running sg.build
> to build a new st.conf, removing the existing /dev/rmt drivers,
> editing /kernel/drv/st.conf and doing a reconfiguration reboot,
> reinstalled the sg drivers with the sg.build and sg.install commands
> and run the device configuration wizard in NetBackup to detect the
> drives. The drives are properly detected, but they don't work. Below
> is an example of what happens when I try to load a tape with robtest,
> note how the drive can correctly return the barcode of the loaded
> tape, but as soon as I try to unload the tape the command fails with a
> 'media not present' error. Any help that anyone can give me in
> resolving this would be greatly appreciated. These drives are less
> than three months old so I have a hard time believing that I've had
> two of them fail at the same time.

> READ_ELEMENT_STATUS complete
> m s11 d1
> Initiating MOVE_MEDIUM from address 1010 to 500
> MOVE_MEDIUM complete
> s d
> drive 1 (addr 500) access = 0 Contains Cartridge = yes
> Source address = 1010 (slot 11)
> Barcode = K01043

All this shows is that the robot can see the tape in the correct
position. 

> Opening /dev/rmt/0cbn, please wait...
> Error - cannot open /dev/rmt/0cbn (I/O error)

With one tape in one drive, I'd do:

for i in /dev/rmt/*cbn ; do mt -f $i status ; done

Do any of the devices report a tape?  If so, the drive mapping in the OS
is being reconstructed incorrectly, and *that* device is the top unit.

-- 
Darren Dunham                                           ddunham at taos.com
Senior Technical Consultant         TAOS            http://www.taos.com/
Got some Dr Pepper?                           San Francisco, CA bay area
         < This line left intentionally blank to confuse you. >

<Prev in Thread] Current Thread [Next in Thread>