The veritas NT driver does support 9840A FC.

See here
http://support.veritas.com/dsl/lists/nbu_device_ddProduct_NETBACKUPDC_prod1_
NBUDC341_vendor_STK.htm

Reserve/release needs to be turned on (its not on by default with 3.4.3).
I assume you have read the patch readme and enabled it.

To check reserve/release is functioning do the following. While NT is
backing up to the tape drive
execute "mt status" against the same (SSO) tape drive from a Netbackup Unix
Media server.

If you get permission denied or an equivalent message the tape drive has
been scsi reserved.
If the command returns tape drive status as you would expect in a non shared
environment, scsi reserve/release
is not functioning.

A tape device will rewind if it:

a) receives a rewind instruction
b) is reset, via a SCSI target reset or IPL or powered on and off

In an SSO tape environment its advised you disable all HBA's (used for
accessing SSO
tape drives) from sending target resets. By default the qlogic should have
this setting disabled
however you may have other HBA's in the SAN with this setting enabled.

I agree, disabling the position check is not a solution.
The position check highlights that you can not bargain on the integrity of
the data if
the position netbackup expects is not what was reported by the tape drive.

At the end of the day two initiators are attempting to access the same
device at the same
time. It is up to Netbackup SSO to manage arbitration to tape devices
between Netbackup media server.
If you assume Netbackup is managing the arbitration correctly then the other
possibilities are:

1. its a configuration issue, in particular drive binding to netbackup tape
drive name consistency.
2. something else other than netbackup is accessing the tape drives. This is
not permitted by Netbackup
   SSO however Netbackup can not enforce it. Enabling scsi reserve/release
will go some way to
   protecting your backups from rogue applications.

Regards
Peter Marelas

-----Original Message-----
From: mander47 AT sca.csc DOT com [mailto:mander47 AT sca.csc DOT com] 
Sent: Monday, 19 August 2002 6:12 PM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] External event caused rewind during write (UPDATE)


Hi All

I have tried to install Veritas Device Driver for NT, but it dosn't support
our STK9840A FC drives.

I'm not interested in disabling the position_check.

I have not been able to find anything in the bpbrm logs.

As it only occur one of our NT media servers, I'm starting to wonder if it
could be caused by the HBA-settings.

I found the following in the eventlog under ql2200 (We are using Qlogic 2200
HBA's):

The device, \Device\ScsiPort2, did not respond within the timeout period.

Could you tell how I check the NT machine is running with SCSI
reserve/release (I'm a UNIX guy).

Regards
Michael

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu