SAN Backup Fails with I/O Error

[SeanDArmstrong]

Greg,
The switch(s) are at firmware v101G12.

The Fibre adapter drivers were updated to version v5.4.82.16 from the MSA1000 Support CD when I upgraded the MSA1000 firmware from 2.32 to v4.32 about 6 months ago (in an attempt to fix this problem!!)

I have installed the ADU and attached the report. Thanks for your help so far.

[Greg Carlson]

Sean,

The latest driver is now 5-5.10a9 or 10a10 for a large SAN envronment.

Verify here:
http://h20000.www2.hp.com/bizsupport/TechSupport/DriverDownload.jsp?pnameOID=315735&locale=en_US&taskId=135&prodTypeId=12169&prodSeriesId=315733&swEnvOID=181

As far as the ADU output, your internal hdds off the 5i controller ID's 0,1 have recent errors as recent as 1-5 days ago.
Port 1 ID 00 has 8 timeout errors
Port 1 ID 01 has 7 hard reads (both could be causing problems for you)
You may have an issue going on with your hdds where your NOS lives. You have some hdds which are showing errors as well in the MSA, however none of the errors are recent, the most recent errors on the MSA show over 180 days ago.

Ciao,
Greg

[Greg Carlson]

Sean,

Also was any work performed on your shelf that is off SCSI B? That is port 4 and is where you had some bus fault errors. The errors there were about 180 days ago. It wouldn't hurt to verify cabling and checking for bent pins.

Ciao,
Greg

p.s. points are appreciated, I like to see which answers helped and which ones didn't.. :)

[Andrew_168]

Sean,

There may be a couple of causes, one could be a SCSI reset ocouring during the backup, this would be rectified by zoning, also I have had a customer who's entire stock of tapes was ruined by the tape drives, the drives caused tape edge damage (This is a known problem), the drives and tapes were replaced by HP under warrenty, I seem to remember the replacements being called silver standard drives.

If this problem has been going on for a period of time I suggest you try another backup program, just in case. Data Protector is a free download with 60 days instant on license, or you could try Arcserve.

Rgds.

[david_799]

This is a bit like deja vouis, I have got almost the same setup (except I am using a clusterd veritas solution)and am having the same problems (and more, possibly because of the veritas cluster), went through the updating everything (with veritas and HP on the phone) i.e. firmware, drivers, even creating .udl files on the servers checking odbc connectivity, using filemon etc, etc.
I also have the added benefit of the catalogs not being written, sometimes!
Have even got the mobile phone number of one of the veritas 3rd line support guys, whom I phone at home in the evening or weekends if needs be.
Veritas admitted to me about 5 months ago that there was a problem with their product in this type of environment, they only took it seriously when American clients started to suffer the same problems, they suggested that possibly an upgrade to 9.1 would sort my problems out(I have my doubts), they now reckon it definatly will, 1 issue I have is that now version 10 is or has just been released.
3 weeks ago I did an exchange course and 2 other guys from different organisations had almost the same setup as we seem to (I've got an msl 5030 and bexec 9), they have the same issues.
I have learned a few tricks with the BEUtility.exe that seem to resolve issues for a while, like deleting and re-adding the media servers and re-setting the primary SAN SSO Server.
Hope this helps, it's just you have the exact same errors.