Hi,

 We have a fairly similar setup here:  Master and Media on the same server 
running NBU6.0 + MP3.  We have been running MP3 now for 4wks.  The only 
difference is that were on Solaris 9.  I haven't seen any problems like that 
since we upgraded to MP3.

 

Ken Frinzl
E-Mail: Ken.Frinzl at SAS.com
Work: 919-531-1905  Cell:919-368-7072
Strategic Recovery

-----Original Message-----
From: veritas-bu-bounces at mailman.eng.auburn.edu [mailto:veritas-bu-bounces 
at mailman.eng.auburn.edu] On Behalf Of Bhangui, Sandeep - BLS CTR
Sent: Tuesday, July 25, 2006 2:17 PM
To: veritas-bu at mailman.eng.auburn.edu
Subject: [Veritas-bu] Scheduler woes with Ver 6.0 + MP3

Hi
        Wondering whether anyone on this forum has seen this behavior.

Environment:  Solaris 8 shop.

Master and Media on the same server running Netbackup Ver 6.0 + MP3.
This is a fresh 6.0 install.

Adic scalar 24 attached to the Master server.

Backups run fine for few days and than the scheduler just stops working.
The scheduler can be made to work only by stopping and starting Netbackup.

Some features seen in the process are.

1. Typically the scheduler stops working when the backup on one of the clients 
has failed with a status code. From the status code one can figure out why the 
client failed. But I cannot find the reason as to why the scheduler just stops 
working when such a incident occurs.

I think in the normal scheme of things if a backup on a client fails than the 
software should just put the Status error code and move on to the other 
scheduled policies and perform the backups for those policies.
But that is not happening. Once the backup of a client has failed no other 
policies scheduled after that runs.

2. Once the above mentioned situation happens. If I try to run 
"/etc/init.d/netbackup" or bp.kill_all to stop Netbackup. It cannot stop the 
Policy Execution Manager, it just hangs while trying to stop that.

No choice but to CTRL C out and than use kill -9 to kill all nbpem processes. 
Once those are killed and I run /etc/init.d/netbackup stop or bp.kill_all it 
stops all the remaining processes and comes back to the prompt.

Somehow my nbpem process is getting hosed/or hung . I am working with Support 
on this and provided them the requested logs. Not heard back from them and 
hence wondering if anyone has seen anything like this, as I know Ver 6.0 has 
lot of issues with Scheduler but was told by support that all necessary patches 
for scheduler are included in MP3.

Any suggestions.

Thanks
Sandeep


_______________________________________________
Veritas-bu maillist  -  Veritas-bu at mailman.eng.auburn.edu 
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu