[Veritas-bu] Possible Disaster Recovery Exercise Problem introduced in

2009-10-02 14:28:16
Subject: [Veritas-bu] Possible Disaster Recovery Exercise Problem introduced in
From: AT parker DOT com
To: veritas-bu AT mailman.eng.auburn DOT edu
Date: Fri, 2 Oct 2009 14:23:46 -0400
Our last full blown DR exercise was done with 6.0 MP5, and our next one 
will be done with  My master is on AIX 5.3 TL9.

I replicate my NBU server's operating system to another disk subsystem and 
boot the copy on somewhat similar hardware in a remote location. 
Everything works fine in this respect.  Naturally, I must remove old 
definitions for network adapters and tape drives and re-detect them, but 
that always works as advertised.

One of the first things that I do is remove any entries to SERVER and 
MEDIA_SERVER for my media servers in the bp.conf and DEVICE_HOST and KNOWN 
in the vm.conf, so that the master server isn't trying to contact them for 
anything.  Media servers are not brought back in the DR exercise.  They 
simply exist "back home" for speed using the SSO option.  They all contain 
a considerable amount of data, and trying to push that data over the 
network while everything else is backing up doesn't make sense.  For DR, 
everything is a restore, so pushing that data back over the network is 
fine, since it is a one-time-good-deal. 

We did a mini-test in August where I restored one Windows server, but 
noticed that with, everything was slow, but I wrote it off to the 
fact that the servers that are media servers at home, and are not media 
servers in DR, didn't exist during the mini-test, and the master server 
was still trying to communicate with them, because they still existed in 
the EMM database.  But now I am looking at it a little more suspiciously, 
thinking it may be a 6.5 thing, instead of a "lack of hosts altogether" 
thing.  My quick answer in August was "It was trying to communicate with 
servers that weren't there and just timing out, but in January (when the 
big test happens) it will get "connection refused" because the same hosts 
will exist but not be media servers.

Is my logic flawed, or should I be looking to decommission the media 
servers in January as part of the test?  If I have to decommission the 
media servers, it is looking like an additional 2 hours due to the media 
movement I have to do in the database to actually remove the media 
servers.  It takes about 3 seconds a tape to move media, and I have a 
bunch of tapes to move logically, even though those tapes won't be touched 
at DR (I restore from duplicated copies).


I apologize for the long-winded e-mail, but do appreciate your input.




PLEASE NOTE: The preceding information may be confidential or
privileged. It only should be used or disseminated for the purpose
of conducting business with Parker. If you are not an intended
recipient, please notify the sender by replying to this message and
then delete the information from your system. Thank you for your
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu

<Prev in Thread] Current Thread [Next in Thread>