Dear
Gabriel,
Thanks for your quick
reply.
The funny thing is that
the other media server through which it is going fine is also installed today
and doesn’t have any CLIENT_READ_TIMEOUT. The values are all
default.
Another funny thing is
that the media server through which the Informix backup is going fine is giving
the same error 13 for its own Informix backup. This server’s Informix backup is
going fine through first media server’s
storage-unit.
So, basically I
installed 3 NBU6.5.1 media servers. All of them are not able to take their own
Informix backups but are able to take each other’s
backup.
All these media servers
have default CLIENT_READ_TIMEOUT and
CLIENT_CONNECT_TIMEOUT.
I think I will follow
your advise and try to change the CLIENT_READ_TIMEOUT and try a
backup.
Previously, all these
had NBU 4.5 media server software installed and had increased TIMEOUT
values.
Secondly, thanks for
your observation about barcodes. Actually the bar code is 8 characters but media
ID is generated using last 6 characters. I don’t have a MEDIA_ID_BARCODE_CHARS
setting in vm.conf. I will read the technote to understand its
benefits.
Regards,
Neeraj
From:
Rosenkoetter, Gabriel [mailto:Gabriel.Rosenkoetter AT radian DOT biz]
Sent: Thursday, July 31, 2008 9:49
PM
To: Neeraj Puri (DHL MY); veritas-bu AT mailman.eng.auburn DOT edu
Subject: RE: [Veritas-bu] Informix Backup
failing with Error 13
Looks like you need to raise
CLIENT_READ_TIMEOUT (and maybe, but probably not, CLIENT_CONNECT_TIMEOUT; don't
try that unless only READ isn't enough) on the media server trying to write its
own Informix backups. (I presume you had, long ago, raised this value on the
other media server for a similar problem.)
The circumstance in which you'd need
to bump CONNECT_TIMEOUT too is f you've got a lengthy backup_start_notify in
place (to, say, put Informix in backup mode or something like
that).
As an aside, that's a pretty kooky
volser. It looks like you're using LTO-2 media and your robotic device is
passing all 8 characters through to NetBackup (not all can be configured not to)
and you don't have a MEDIA_ID_BARCODE_CHARS setting in vm.conf, but you probably
want one, although addding one now may be a bit touchy, since you've already
inventoried a bunch of tapes with the traily L2, and you'll need to blow the
media and vol DB entries for those away and recreat them, which will definitely
be a mess. (Cf, http://seer.support.veritas.com/docs/236271.htm.)
--
gabriel
rosenkoetter
Radian Group Inc, Unix/Linux/VMware Sysadmin / Backup &
Recovery
gabriel.rosenkoetter AT radian DOT biz, 215 231 1556
From:
Neeraj Puri (DHL MY)
[mailto:neeraj.puri AT dhl DOT com]
Sent: Thursday, July 31, 2008 9:04
AM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] Informix Backup
failing with Error 13
Dear
All,
I installed a NBU 6.5.1 media server
on a HP-UX box. The File System backup of this box on its own storage-unit works
fine whereas the Informix backup fails with Error 13. Here is the screen of shot
of the exact error.
07/31/2008 19:54:21 - started
process bpbrm (pid=1068)
07/31/2008 19:54:29 -
connecting
07/31/2008 19:54:36 - connected;
connect time: 0:00:00
07/31/2008 19:54:43 - mounting
0148L2
07/31/2008 19:55:38 - mounted
0148L2; mount time: 0:00:55
07/31/2008 19:55:38 - positioning
0148L2 to file 1
07/31/2008 19:55:45 - positioned
0148L2; position time: 0:00:07
07/31/2008 19:55:45 - begin
writing
07/31/2008 19:58:35 - Error bpbrm
(pid=1762) socket read failed: errno = 52 - Stream ioctl
timeout
07/31/2008 20:00:58 - end writing;
write time: 0:05:13
file read failed
(13)
If the same Informix backup is sent
through another media server, it works fine.
Please advise if you faced similar
problem or know the reason for this failure.
Regards,
Neeraj