Veritas-bu

[Veritas-bu] NB 6.5.1 issue, error 25, jobs fail

2008-03-12 13:57:14
Subject: [Veritas-bu] NB 6.5.1 issue, error 25, jobs fail
From: "S Conn." <sysadminlists AT gmail DOT com>
To: Veritas-bu AT mailman.eng.auburn DOT edu
Date: Wed, 12 Mar 2008 12:27:09 -0500
I had Netbackup for a few years now, 4.5, 5.1, 6.0 MP4.  I finally got
a new backup server and decided to fresh install NB 6.5.1 and start on
it from scratch.  I didn't import the catalogs or anything over.  I'm
running Windows 2003 Enterprise R2 SP2.

Well, after battling I got it working, except for one major issue.  It
will back up for a day or two or sometimes three, then it will have an
issue.  I'll come in the next morning to find that half my jobs
failed.  I try to launch my Admin Console I get an error message:

Unable to connect to the selected NetBackup host "bkpsrv01".
1. Make sure the user has privileges on the host.
2. Make sure the local host is listed in the server list of the
destination host "bkpsrv01".
3. Make sure there is a valid network connection.
4. Check authentication.
5. Make sure all the services are running on the host.

If I hit ok, it gives me a "Change Server" dialog, asking me to choose
a different NB server.  I only have one.  Rebooting the services
doesn't seem to fully help, rebooting the server itself does.  And
then it will work fine for a few days.  When the issue happens, all
services are running.  I restarted services one at a time to
troubleshoot the issue.  Restarting the "Netbackup Client Service"
worked but I was still getting some errors while working in the
software.  A reboot clears everything up just fine..

See here for the first error:
http://i207.photobucket.com/albums/bb110/netmancer/SysAdminHell/1stNBerror.jpg
See here for the second error:
http://i207.photobucket.com/albums/bb110/netmancer/SysAdminHell/2ndNBError.jpg

Looking through my logs, I'm seeing everything running fine, writing,
etc, but then I suddenly get an error "db_FLISTsend failed: cannot
connect on socket (25)"  After that, nothing else will write.
Anything already writing will quit, then all the new jobs will queue
up and eventually time out without even attempting to write.  There's
no other error that I can tell that caused it.  Even the Windows event
logs don't show anything significant at that time.

I've searched around without much luck.  Does anyone have any ideas or
has anyone seen this problem before?

Thanks,

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