Cleaning drives of DLT 8000 2/20

[Paul Bout]

Altough Omniback 3.50 II on HP-UX 11.0 is configured for automatic cleaning one of the two drives of a DLT8000 robot nothing happens when one of the two drives gets dirty.
The problem would be solved after applying patch PHSS_24428 and upgrading both the firmware of the drives and the library to the latest version but unfortunately it didn???t.
The testcase as mentioned in the Omniback manual works fine.
The only way i can detect a dirty drive is via the HP Tape Library Report Summary in my web browser but i rather do that automaticaly.
Doea anybody has the same experience?
Thanks in advance,
Paul Bout

[BR19682]

Hi there.
Did you setup one slot as slot for cleaning tape ?
There maybe a file named dirty... something.
This may be another reason. If you find it in one of the OmniBack directories, you should rename it.
Rgds
Alexander M. Ermes

[Paul Bout]

No files named dirty on the cell server. They are only used when you want to test.
Setting up a slot for a cleaning cartridge in Omniback or in the library itself ????
In Omniback i configured a slot for a cleaning cartridge but i can't find a setting for defining one in the library itself if that is necessary at all.
Greetz, Paul

[CA713937]

You might want to check the Omniback cleaning logs, to see if cleaning was actually performed on the dirty drive.

On Windows: \log\cleaning.log

On HPUX: /var/opt/omni/log/cleaning.log

[Paul Bout]

Hi David,

Yes, i did that too. No entries in the cleaning logfile.

Cheers, Paul

[BR19682]

Hi there.
Sorry for being a bit late.
Did you check the dirty drive detection button in the drive definition of OmniBack ( advanced options ) ?
Rgds
Alexander M. Ermes

[Paul Bout]

Hi Alexander,

Yes that checkbox is selected too.
I think Omniback didn't get a sign that a drive is dirty from the tape robot. So i wonder where i can configure that if that possible at all.
Greetz, Paul

[CA628575]

Your assumption that OB may never have received the cleaning request is reasonable.

There are two ways that a drive can notify the software of a cleaning request.

a) In response to a read/write command the drive can respond with a check condition (ie error code) that cleaning is required.

This usually happens in the more drastic cases.

b) The drive can set a flag (which is called a Tape Alert flag) that shows that cleaning is recommended. This is usually done in less critical cases, as the application does not have to respond immediately and can perform the cleaning after the current operation is completed.

However, case b) does require the application to monitor TapeAlert flags. I assume OB does this, but I'm not entirely certain at the moment.

There could be an explanation for your symptoms in case b). Tape Alert flags have the interesting twist that they reset on read. Thus if you have more than one software package reading and monitoring Tape Alert flags it is non-deterministic which of the two applications gets the flag. Thus it is possible that the drive attempted to notify the application, but someone else saw and cleared the notification first before OB had a chance to react.

Are you running any other tape monitoring software, hardware monitors, etc.?

[Paul Bout]

Jan,

Concerning your story about the tape flags i fully agree. Unfortunately no other monitoring tool is runnning.

Greetz, Paul