Chris,
What I have done for similar DR tests (when recovering the entire
catalog) is edit the hosts file on the master server and add the media servers
as aliases to a master or media server that you will be running at the DR site.
In the end, NBU is trying to communicate with the media servers and/or resolve
certain hostnames via DNS and is timing out while doing so and causing your
performance issues. Also, I find it
helpful to scan the bprd and bpcd logs looking for hostnames that NBU is trying
to find/resolve and then adding them in there as well.
If nothing shows up in those logs and you are still having performance
issues, you can also check for rogue hostnames in the Vx logs via: vxlogview –p 51216 –t
00:10:00
Sample
hosts update:
/etc/hosts
10.10.1.1 nbmaster nbmaster.domain.com nbmedia1 nbmedia2 nbmedia3
If that doesn’t fix it, then run the “nbemmcmd –listhosts” command and add the master or media server
hostname as an alias for the media servers and ndmp hosts listed.
nbemmcmd
-machinealias
Usage:
nbemmcmd
-machinealias [-brief] [-emmname <string>] [-emmport <unsigned
integer>]
[-addalias -alias <string> -machinename <string>]
[-deletealias -alias <string>]
[-deleteallaliases -machinename <string>]
[-getaliases -machinename <string>]
-machinetype <api | app_cluster | client | cluster | disk_array | master |
media | ndmp | server | virtual_machine>
HTH
-Mike
From:
veritas-bu-bounces AT mailman.eng.auburn DOT edu [mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu]
On Behalf Of Christopher Costa
Sent: Friday, June 11, 2010 1:13 PM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] Catalog Recover - DR
Hello to all,
I have a current situation that I need assistance on and I
am sure someone on this DL can help me with.
We current have a 6.5.1 NBU environment with a Solaris
Master Server and 6 Media Servers (a mixture of AIX, Solaris and Windows)
When we perform a DR recovery test, we take the catalog tape
and perform a Hot Catalog Recovery using the DR file, etc. The recovery of the
catalog DB works fine, the issue is that in our DR site we do not have the 6
media servers. As we all know the Master Server will respond poorly if it
cannot communicate with the media servers configured in the EMM database and
bp.conf file.
We run the Symantec recommended commands to remove the media
server entry in the EMM database and Volmgr DB, as well as comment out these
media servers in the bp.conf.
However these media server entries still show up in the EMM
database and no matter how hard and how many times we run these tasks it never
works correctly.
At this time we are failing in our audits (RPO/RTO is
24-hours) due to the responsiveness of the Master Server, while it continues to
search for these media servers.
Symantec says that we need to have these 6 media servers in
the DR site and does not support the methodology we are following.
Does anyone have any ideas?