StoreEver Tape Storage
Showing results for 
Search instead for 
Did you mean: 

Data Protector: Major Error 90:51

Smeyers Arne
Occasional Advisor

Data Protector: Major Error 90:51


I have an MSL6030 with two LTO-2 tape devices connected by FC. I have about 10 different back-up jobs which run during the week. Only one back-up jobs just drives me mad. This jobs consists a full back-up of 6 different servers. The back-up starts, but after a half-hour it gives me this message:

[Major] From: BMA@ "Drive 1" Time: 20/10/2006 9:25:08
[90:51] Tape0:1:1:1
Cannot write to device (Details unknown.)

My tape is set to Poor and the back-up job hangs. At that time, abaout 10 to 25 GB has been back-upped.

I've tried to perform the same job on new tapes, same result. When I change the block size, same result. I change the concurrency (from 4 to 1), same result.

When I split up the job into smaller jobs, everything works fine. But the thing is, this jobs exists for one year now, and it just gave me this errors since a couple of weeks.

Friendly greetings,

Peter Godron
Honored Contributor

Re: Data Protector: Major Error 90:51

Hi Arne,
and welcome to the ITRC forums!

Only two things I could think of straight away:
1. Faulty cable
Solution: Replace cable
which may lead to:
2. Timeouts on write
Add OB2SPTRETRY and OB2SPTSLEEP in .omnirc file
Defaults are 10 retries and/or 10 seconds

If these answers were helpful, please read:
Smeyers Arne
Occasional Advisor

Re: Data Protector: Major Error 90:51

I've tried this, but same result.

More info: I use DTP 5.5... In case this is important?

Frendly greetings,

Bill Costigan
Honored Contributor

Re: Data Protector: Major Error 90:51

First off you are using Tape0 which means that the windows tape drivers are enabled. These should be disabled. DP can either use the tape drivers Tape0:1:1:1 or the raw scsi drivers scsi0:1:1:1. With the tape drivers enabled, windows (and all the software running on the severs) know these are tape drives and might decide to access or configure these.

I've had this happen in four different situations.

1. another backup software product (arcserver) was running on a server with access to the libary through the SAN. It did a device discovery, found the drives and executed a SCSI reserve on the tape drive during the DP backup.. DP lost complete access to the drive.

2. Some windows media management software, noticed that I mounted a tape in a drive (a DP formated tape) and decided what I really needed was an new label written over the the DP header.

3. Some system monitor tool (lighthouse, I think) woke up every 30 minutes and noticed that there was a tape drive that really needed a status check performed on it. BTW the status check aborted any pending I/Os including the DP backup that was running from another server.

4. An older version of compaq's insite manager decided that the I/O timeouts for my tape drives should be 90 seconds even though DP set them for 30 minutes. So if DP did any operation (e.g., scan to the end of tape so I can append) and the operation took more than 90 seconds, the IO aborted.

In all these cases DP had no idea what happened to it and marked that tape as bad after reporting the error.

The bottom line is you need to be careful what other softeware has access to your taped drives if you library is SAN connected.