[Veritas-bu] Are you using a Powderhorn?
2002-01-23 15:11:11
Subject: |
[Veritas-bu] Are you using a Powderhorn? |
From: |
David A. Chapa" <david AT xbpadm-commands DOT com (David A. Chapa) |
Date: |
Wed, 23 Jan 2002 12:11:11 -0800 (PST) |
Then take note. (FYI: This is documented in the
technote I mention below...however, this email is more
a request for an enhancement under 3.4.x than it is a
bug report).
And Veritas Engineering, please do respond either to
the list or me individually...its okay I used to be
with OpenVision, you can talk to me.
Here it goes:
I am sitting at my client site as I type this after
seeing the following.
They have a big STK Powderhorn with 5000 tapes using
ACS/LS. They also use DISABLE_COUNTMEDIA (see
http://seer.support.veritas.com/docs/237534.htm
for more information) because of the large volume of
tapes in the library the process (countmedia) took
forever.
However, when the volume pool runs out of available
media and none are in the scratch pool, bpsched -
mainempty invokes bptm -countmedia again, overriding
the DISABLE_COUNTMEDIA that we set up. Now I know the
tech note says that this may happen under the
conditions previously mentioned, but this has been
known since July of 2001 (maybe earlier) and seems to
be a pretty big deal, especially at these big sites
that backup 3 to 4TB of data PER DAY incrementally.
They go through tapes like water around here.
Anyway, what happens is the bpsched processes that are
looking for more media through bptm eventually become
defunct. During this fiasco, you can't query anything
in your Media Manager Database because vmd is so tied
up with these count media requests that it returns a 39
or 70 error. I was able to terminate vmd after several
attempts and slowly the bptm processes died off.
however, we weren't able to get NetBackup to respond
beyond that and had to reboot our server (to clear the
defunct processes and get NBU back on its feet.) This
issue leaves tapes stranded in drives as well, which
NBU knows nothing about after the reboot.
My question is this:
Veritas Engineering is there going to be a fix for this
under 3.4.1? Or will they be forced to move to 4.5
when that comes out? And if that's the case, will it
in fact be fixed in 4.5?
One final note: I'M NOT BASHING VERITAS...AS A MATTER
OF FACT KUDOS TO VERITAS FOR THE WORK THEY HAVE DONE
DOCUMENTING THIS AND OTHER ISSUES FOR US IN THEIR
KNOWLEDGE BASE.
Thanks
David
<><><><><><><><><><><><><><><><><><><><>
David A. Chapa
Consulting Manager
DataStaff, Inc.
847 413 1144
http://www.consulting.datastaff.com
---------------------------------------
http://www.xbpadm-commands.com
NBU-LSERV AT datastaff DOT com - Adv. Scripting
|
|
|