DLT80 (Int) drive bad I/O error; drive stops responding

[Joe Nalducci]

OK I have been using a HP sure store Quantum DLT80Int drive for over about 2 years now, never had a problem. It connected to an Adaptec 2940U2 . Then one day while changing tapes I heard the infamous Groaning noise. The tape was stuck in the drive for a bit then was released. After this event strange things started brewing.
With no changes to software or hardware on a Windows NT 4.0 SP6a, running Veritas Backup Exec I now constantly have I/O errors that cause the drive to go OFFLINE and requires me to reboot the Server almost everyday to get the drive to respond again only to perform a backup job, then an I/O error causes the drive goes offline again.
I get 2 distinct Events in the event log:

Event ID: 11
Source aic78u2
The driver detected a controller error on \device\scsiport1

Event ID: 15
Source qntmdlt
The device, \device\tape0, is not ready for access yet.


When this happens the drive is still identified by NT 4.0 in the Tape drive control panel application with the driver loaded. The Adaptec SCSI card does not disappear.

Self-test says the drive is fine.

Has anyone seen this behavior before or have any suggestions before I rip apart this machine and replace the SCSI card and or DLT drive?

Thanks

[Andy Bloor_1]

I had a similar issue with DLT8000 drives in a SureStore Library (although SAN attached). Drives going offline, timing out on jobs, I/O error mid backup etc.

The problem started with one drive, and then slowly moved onto all drives.

After much testing, upgrading firmware etc etc, the only thing left was to replace the drives (although it took a lot of convincing HP engineering to do this as there was no 'hard' error).

We changed both drives and all has been well ever since. This problem then arose an another site within a week or so (both libraries were purchased at the same time), so we changed the drives there too.

The only thing I can see we were doing which is not good practice was cleaning the drives every Friday prior to the weekly backup, whether they needed them or not. I have read since that this is not good practice, and DLT's should only be cleaned when required (wears read/write heads I understand).