Well NetBackup 6 includes support for sharing drives between NDMP and 
non-NDMP use, so must include logic that probably was not there before. 
Check the version of ONTAP is supported with your version of NetBackup. 
You can turn on detailed NDMP tracing on the "NetBackup for NDMP Server" - 
where the option is installed.   It gets very bulky but can trace every 
NDMP exchange.   If your clocks are synch'd you may be able to see what is 
happening.

The IBM LTO-2 drive does support the "WRITE ATTRIBUTE" command, according 
to the manual.  It writes to the "Medium Auxilliary Memory" i.e. the 
LTO-CM cartridge memory.  I guess it's possible that NetBackup 6 now has 
support for attempting to write to this, as one field available for use is 
"Backup Date".   Maybe this is not working properly - without the detail 
of what was being written you cannot see if it is being done correctly.

William D L Brown




"George Drew" <gdrew AT deathstar DOT org> 
Sent by: veritas-bu-admin AT mailman.eng.auburn DOT edu
15-May-2006 19:07
 
To
"Len Boyle" <Len.Boyle AT sas DOT com>
cc
veritas-bu AT mailman.eng.auburn DOT edu
Subject
Re: [Veritas-bu] Strange errors in netapp messages log after upgrade of 
nbu from 5.1 to 60MP2 -- diag gear






Len,

This message means that you have some device in the path to your
tape drive that isn't happy with 16-byte CDBs (some scsi-3 commands).
This one in particular is the scsi-3 "WRITE ATTRIBUTE" command.

(Reference: http://t10.org/ftp/t10/drafts/spc4/spc4r04.pdf)

George

On Sat, 13 May 2006, Len Boyle wrote:

> Good Day,
>
> We upgraded a netbackup master running on solaris 9 from 5.1mp3 to 
60MP2. This master runs ndmp backups on 6 R200's with fibre direct 
attached IBM LTO-2 tape drives. After the netbackup upgrade, with no 
changes with the tape drives, tape library or netapps we started to see 
the following messages in the /etc/messages file on the netapp. We are 
being told by netapp and symantec that this is a problem with scsi 
commands, but they report that they can not tell us what the offending 
scsi commands are, unless we hook up a sniffer.
> So my question to this group, is Any hints on what one should  use to 
act as a sniffer for this scsi connection?
>
> That is what company makes good sniffers, and what are the different 
features  one should look for?
>
> Thanks for any info you guys can throw this way.
>
> Fri May 12 07:54:31 EDT [scsi.cmd.contingentAllegiance:error]: Device 
3a.51: Contingent allegiance: cdb 0x8d.
> Fri May 12 08:00:01 EDT [kern.uptime.filer:info]:   8:00am up 57 days, 
18:59 427592 NFS ops, 3947850238 CIFS ops, 9 HTTP ops, 0 DAFS ops, 0 FCP
> ops, 0 iSCSI ops
> Fri May 12 08:03:16 EDT [scsi.cmd.contingentAllegiance:error]: Device 
3b.52: Contingent allegiance: cdb 0x8d.
>
> _______________________________________________
> Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu