I/O bus failed to reset

[penauroth]

I have had Veritas Backup Exec v8.6 installed on our W2K Dell PowerEdge 2600 server since 11/2002. Backups are done using a Dell PowerVault 110T DLT VS 80 tape drive. I have incremental backups (no overwrite) running Monday - Thursday and a full backup (overwrite enabled) on Fridays. Also, I have 4 sets of tapes - one for each week of the month. Over 50% of my backups fail with the error message: "The I/O bus failed to reset." Sometimes bytes are written to the tape but most of the time 0 bytes are written. This problem has occured ever since we installed it on the server. Within the past month, it has gotten worse. No applications or Windows updates have been installed or applied recently. The following is a sample job log:

======================================================================
Job server: PHNXSERVER
Job name: Sandbox
Job started: Tuesday, May 13, 2003 at 11:00:05 PM
Job type: Backup
Job Log: BEX01.txt
======================================================================

======================================================================
Job Operation - Virus Scan
Attempt to clean infected files enabled.
Infected files will be skipped.
======================================================================

Scan of "D:".
Virus Scan started on 5/13/2003 at 11:00:05 PM.

13 files were scanned for viruses.
0 files were infected with a virus.
----------------------------------------------------------------------

Drive and media information from media mount:
Robotic Library Name:
Drive Name: DELL 1
Slot: 0
Media GUID: {C09541D9-24E3-4964-B593-C1F013559966}
Media Label: Tuesday Tape 4
Overwrite Protection Time Left: None
Append Time Left: Infinite
Targeted Media Set Name: Phoenix Backup

======================================================================
Job Operation - Backup
Media operation - append.
Hardware compression enabled.
======================================================================

Storage device "DELL 1" reported an error on a request to space to end of data.

Error reported:
The I/O bus was reset.

^ ^ ^ ^ ^



======================================================================
Job ended: Tuesday, May 13, 2003 at 11:03:30 PM
Job completion status: Failed
======================================================================

Full backups are fine. I have checked Veritas knowledge base and cannot find any articles related to this issue. Thanks in advance.

Paul

 

[Rols]

I/O bus reset means that the SCSI bus between the server and the tape device was reset.

Check all cable connections (might be wise to replace SCSI cable & terminator just to be sure - also check that the pins on the SCSI connectors arent bent).

Check that you have the latest drivers for your SCSI card
(You should also use the Veritas drivers for your tape device)

If all else fails replace the tape device (or at least test another tape device on the server). Also try and do a backup using NTbackup and see if you get any errors.

Cheers

Rols

 

[tonyhake]

Penauroth - Did you ever resolve this issue? We have two PE2600's with the same tape drive and BackupExec and experience the same error message. Please let me know, I would appreciate your feedback.

Tony

 

[penauroth]

Haven't got around to it yet but I will and record my findings here.

Paul

 

[SQLBill]

Have you rebooted the backup server (the one with BE installed)?

-SQLBill

 

[tonyhake]

In my case, just this morning I was trying to restore around 5GB worth of data. About 3GB into the job, it came up with the error message. Rebooted the server, same problem. Tried multiple times with the same thing. Stuck in a cleaning tape, let it run, and coincidence or not, the restore job went fine. All drivers are current.

 

[wtrdog]

We have about 75 Dell PE 2500 / 2600's and ran into the same problem with the Dell TBU's. Turns out this was an issue with the firmware on the onboard controller. It could not handle being a RAID/SCSI Controller. The problem was resolved when Dell sent us some new controller cards for the TBU's only. We went round and round with Dell and Veritas.. but this was the result that fixed us.

 

[penauroth]

I contacted Dell but they wanted a "software contract" to discern the problem. I wasn't about to pay a couple hundred for that.

Veritas stated that I should use their drivers instead of Dells'. So I uninstalled the application and re-installed on another partition and installed the drivers from Veritas. So far, no failures however only two days of backups. I hope this fixes the problem. Thanks for everyones help.

Paul

 

[penauroth]

By the way, I would be happy to email Veritas' document. Just post a request here.

Paul

 

[tonyhake]

penauroth, if you could please email me the doc, I would appreciate it. I continue to have issues with these dang things. If the driver idea doesn't work, I will just call and lie to Dell and tell them the TBU is dead and see if replacing it works.

Thanks!

Tony
tony@codoor.com

 

[SQLBill]

Tony,

Apparently you haven't dealt with DELL on hardware issues. When they send you a new part, they require the old one back within 7 days. If they don't get it back they charge you for the new one until they get the old one back. Then they test/examine the old one and if there's nothing wrong with it they can charge you for the new one.

-SQLBill

 

[tonyhake]

Thanks for the warning, Bill. Actually deal with Dell extensively on hardware issues (all our servers and workstations are Dell) and have never had a problem.

Usually I just tell their support people what steps I followed to troubleshoot (it helps to look at their web site's support stuff first as that is what the tech follows) and then tell them what I think is wrong and I get it. They have never charged me back for a part - of course, IMHO, I am rarely wrong. Haha.