Major Problem, Backups downing Drives!

[hatman]

Hey Guys, I've got a major problem that just started yesterday. Every time I try to start a backup (scheduled or manual) it downs the drive. (we've got 2 drives and it downs 1 then 2). I can up them, but it does the same thing, the backup never actually starts, just is out there in limbo. If I resubmit the request,or do the backup again, the drives are downed again. Detailed status only shows Tape Mounting.
The only thing in the problems report is scheduler exitiong - no storage units availabl... which is because the backup downed the drives! I've tried cleaning the drives, resetting them, even powering off/on the library..no luck. Please let me know if anyone has any ideas out there. Thanks in advance

 

[comtec17]

Need more info....

What type of libary are you using and how is it connected to the server?

Is this the Master Server or a SAN Media Server?

Are you using SSO?

What type of tape drives are you using and how are they connected?

Which tape drive driver are you using?

 

[hatman]

I've got an IBM 3583 ultrium tape library connected via SCSI cable directly to the Master/Media Server.

No sso or san environment.

2 lto drives within the library with firmware update 25D4.

It just happened yesterday when I came in to work. Drives are downed as soon as the tape is mounted. (actually in Netbackup the tape is always in a 'mounting' stage) According to the lcd on the tape library, the tapes are already mounted.

 

[comtec17]

Can you use robtest and mount/dismount and move media around the library?

 

[sdeb]

Looks very strange in a "simple" SCSI DAS configuration...

You cloud maybe take a look at SCSI ID's, if they have been changed by someone else.
You can also try to do a new "Device Discovery Wizard" and check if Netbackup is able to discover 'em again.

I don't know the Master server OS but you should also check device recognition by this OS...

 

[hatman]

I couldn't connect with robtest at all. So I had to rescan the drives, and library, and then reassign them. Then I was able to use robtest. I'm still not sure why netbackup didn't give me any info on my error reports, but I just got done doing a backup and everything seems ok. I think I've got a shotty scsi cable for 1, but I'll continue to monitor it. Thanks everyone for your help!

 

[birky]

Hi

Seems obvious I know but, have you tried to clean the drives (or should I say recalibrate).

We've had problems with drives going down and a simple clean fixed the problem.

 

[birky]

doh..sorry just reread and saw the bit about cleaning the drives

 

[young9zh]

I think the tape drivers' SCSI ID have been changed and their device name on OS have been changed too. You should reconfigure the devices in NetBackup again.

 

[jdespres]

Check what tapes are involved..... Check the following:
do a grep for "DOWN drives" in the following dir:
/usr/openv/netbackup/db/error

 

[msosacruz]

Any idea for this case?

 

[kurtvonage]

I had a similar problem, but this was because i updated my IBM servraid controllers BIOS to v7 and it started to send resets to the tape drives. The controller is apparently not supported by IBM for attaching to Tape drives (even though it has LVD scsi ports). I switched to a "dumb" scsi card (Adaptec Ultra160) and never had this resets again.
My config is Linux Media/Master server attached to a Qualstar 8466 Library with Scsi LTO2 drives.
Bottom line , ask yourself if you did a firmware/bios upgrade on your raid/scsi controller ? are drives attached to this controller (internally) and did you upgrade the kernel version or OS version for this system (again i am using Linux so i do not know how this affects you)