[Veritas-bu] Hung Jobs

Williams, Kristopher L kris.williams at hp.com
Tue Mar 29 18:05:09 CST 2005


When one of the backups is hung, run bpps on the Win2k3 server. Look for
processes named bpfis. This is the Veritas Snapshot process that
attempts to backup open files. The only way I was able to solve my
problem long term was to disable VSP on all of my clients.



-----Original Message-----
From: Magilsen, Jack (Bolton) [mailto:jmagilsen at husky.ca] 
Sent: Saturday, March 26, 2005 11:44 AM
To: Williams, Kristopher L; veritas-bu at mailman.eng.auburn.edu
Cc: Haime, Terry (Bolton)
Subject: RE: [Veritas-bu] Hung Jobs

Hi Kris

I am having similar problems since upgrading to 5.0 MP1 on a windows
2000 backup server .

So far I have two 2003 Windows clients that hang .

The job starts, the tape mounts, the tape file skips to the file mark
needed .
The client begins to push out the data and then client network activity
stops .

The data rate starts at a good rate then drops eventually to 0 KB/sec .
If you kill it soon enough the job will abort . But if you wait to long
the job is hung with no errors and requires restart of services .

This some times leaves the tape stuck in the drive requiring a hard
reset of the drive .

Most of my other backups run fine .

Still researching the problem ..

Jack ..

-----Original Message-----
From: veritas-bu-admin at mailman.eng.auburn.edu
[mailto:veritas-bu-admin at mailman.eng.auburn.edu]On Behalf Of Williams,
Kristopher L
Sent: Wednesday, January 05, 2005 11:25 AM
To: veritas-bu at mailman.eng.auburn.edu
Subject: [Veritas-bu] Hung Jobs

Hello all,

I am having a problem with jobs slowing down through the backup window
and eventually coming to a point of being hung. The jobs start running
at decent speeds, then trickle down to about 17 - 25 kb/sec. It's
different clients each night, and out of 500 jobs, there might be 7-8
hung when I check the morning after. I am unable to kill the jobs via
activity monitor, the only way I can get them to die is to stop the
services on all media managers and restart them. After this is done the
hung jobs go away and end with status code 23. Has anyone seen anything
like this?

I'm running NetBackup v5.0 MP3 on a Windows 2000 Master server. All
media servers are Windows 2000 also. The library is a STK L700 with FC
LTO Gen1 drives. All drives and robot are at the current level of
firmware. Clients are all Windows based (NT 4, Win2k, Win2k3). All of
the hung jobs are network backups. Because it is different clients each
night, I do not think it is 100% a network problem.

Thanks for the help!


Veritas-bu maillist  -  Veritas-bu at mailman.eng.auburn.edu

More information about the Veritas-bu mailing list