Veritas-bu

Re: [Veritas-bu] NBU and Clustered MSCS fileserver, Flashbackup?

2009-07-30 18:52:19
Subject: Re: [Veritas-bu] NBU and Clustered MSCS fileserver, Flashbackup?
From: Ed Wilts <ewilts AT ewilts DOT org>
To: "Spellacy, Sean" <Sean.Spellacy AT viha DOT ca>
Date: Thu, 30 Jul 2009 17:48:59 -0500
On Thu, Jul 30, 2009 at 3:14 PM, Spellacy, Sean <Sean.Spellacy AT viha DOT ca> wrote:
 
I have recently added  to my job list a Fileserver cluster made up of four win2k3r2 machines with roughly 5TB of data on them, millions of files.
As expected this takes a relative eternity to backup.
 
I was hoping to use flashbackup to be able to pull them but I read in the documentation that flashbackup is not supported on clusters.
 
Is this true? This does not work or this is not supported? Is anyone using flashbackup on win clusters?

It's absolutely false - it does work and we use it a lot on clusters with far more than 5TB of data.

You need to create a snapshot volume and you MUST add this to the cluster group with the data volume.  Then use the cluster virtual server name that contains these volume to do the backup.
 
So far I have broken these jobs down into multiple streams and that has bought me some ground, but I suspect the real issue here is the millions of files being parsed. I have also balanced to jobs across two media servers.
 
Does anyone have any insights or opinions as to how I may be able to speed up these jobs.

We found roughly a 2x performance increase using FlashBackup although I haven't benchmarked it in several years.  Not only that, but the CPU hit on the client is considerably reduced from walking the file systems.  64-bit clients really, really help.

One thing to watch out for is the maximum job on the client.  Because you're using virtual server names, you may end up running multiple jobs on the same physical server.  Although I've complained to Symantec many times about this, they have yet to give us any ability to control the number of jobs you can run.  What I do is dedicate a DSSU to the cluster for the large data volumes and set the limit to 1.  It ensures that only job runs in the cluster but my admins whine like you wouldn't believe if I kick off 2 jobs on the same node.  There's currently no way to ensure that only 1 job can run per node in your cluster.  This is one of the reasons why I have claimed - and continue to claim - that NetBackup doesn't really support clusters.  It tolerates them but real cluster support isn't there yet.

   .../Ed
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
<Prev in Thread] Current Thread [Next in Thread>