I have been having backup issues for a long time and I need to get my backups working 100%.
Here is the equipment that I am using:
Server - HP ProLiant DL380 G8
OS – Windows Server 2008 r2 (fully patched)
Tape Drive – HP Ultrium 5 / T24 Tandberg Tape Library (latest Firmware)
SAS CARD / HBA - HP H221 (no updated Firmware at the moment) (newer card)
SAS CARD / HBA - HP SC08e (no updated Firmware at the moment) (older card)
SAS Cable – StarTech.com External Mini SAS Cable (SFF-8088)
Software – Symantec Backup Exec 2012 r2 (fully patched)
I can get a backup to work maybe 1 out of 10 times, most of the time, I get this error:
“The job failed with the following error: The device timed out.”
Or I get this error:
The driver detected a controller error on \Device\RaidPort1.
This error signifies controller errors. These errors are generally caused by hardware that is not functioning correctly. To resolve these errors, try the following actions:
- Slow the SCSI bus.
- Verify that the latest SCSI drivers and firmware are loaded.
And then Backup Exec turns my tape drive offline.
For example, last night I did a test backup of a 700mb ISO file and it worked fine, then this morning I tried to run the same job and it failed.
I have tried using a different SAS cable and that produced the same results.
I have tried using new tapes and produces the same results.
I can use Backup Exec to back up files perfectly fine to my NAS device. So I don’t think it is a Backup Exec issue.
Symantec support say’s it is a hardware issue with either the tape drive or the SAS Card / HBA.
Tandberg support doesn’t know what to do since it works occasionally, they think it's a server / SAS card issue.
HP suggested that I try a different SAS Card, so I went from HP SC08e to HP H221. I switched out the cards and was then able to maybe do 5 backup jobs in a row to tape 100%. Then on a Friday evening it stopped working and I’m getting the same time out / slow the SCSI bus error. I'm now on my 4th SAS card.
I have even tried using the SAS cards on a different server and that produced the same results.
Does anyone have any ideas?