false hardware error message

[wayne_qcc]

Running BE 11d on a Windows 2000 server using a dual LTO-3 drive Overland Neo 2000 via iSCSI. 20 servers backed up full over the weekend, differential each week night.

Worked like a charm for three months, then on Friday 7/27 early morning, backups started failing with "hardware error" messages. Cleaned drives, no effect. Rebooted everything, no effect. Disconnected iSCSI, reconfigured the connection, no effect.

Ultimately, discovered that the failures seemed to happen when tapes were loaded in both drives (even if one was inactive), so disabled one drive in the Overland and backups started working again. Doesn't matter which drive is disabled, so it's unlikely to be an actual "hardware" problem. Library Expansion Option is licensed and the license is installed (obviously, since it worked for 3 months). No other tape drives configured. Ran Tools/License Keys and Installation again to try to kick-start the LE Option, no effect.

Solution? Other than wiping, reinstalling (including all the Option licenses), and rebuilding all 40 backup jobs from scratch?

[Ben_L_]

Wayne,

Do you have partitions setup for the library?  If so write down how you have the partitions setup.

1. Disable both tape drives in BE, then delete them.

2. Disable the library and delete them. You may get prompted to point jobs at a different devices at this time, just point them at all devices.  You will need to go back later and change them back to the correct device.

3. Cycle the services

4.  When the services come back up all your devices should be back.  Repoint your jobs at the correct device and attempt to run the jobs again to see if the problem comes back.

Also in even the event logs have you seen any errors coming from the SCSI card or library?

Regards,

[wayne_qcc]

Followed your instructions and ultimately got successful backups (Differential) on the 19 servers scheduled overnight and a failure on the one SQL server that's backed up this way. The SQL was an error in deleting the logs snapshot, so I'm not going to worry about that right now; I haven't checked, but I believe all the data was actually backed up.

On the 19 Differentials, though, every job that went to HP 1 (the first tape drive in the Overland) succeeded on the first try. Only two that went first to HP 2 succeeded; 9 failed (including some repeats) until they switched to HP 1, where they succeeded on the first retry.

I've looked through the media server event logs and found nothing related except in the application logs, which basically just repeated the information that a backup had failed. The Overland itself has never reported any errors, not even a need for cleaning (though I did a clean on a hunch after this all started). Communication is via GEO card in the Overland and iSCSI, and that has been rock-solid from the start.