SUMMARY: Veritas Netbackup agent on a TruCluster 5.1 system

From: Goetzman, Dan <Dan_Goetzman_at_bmc.com>
Date: Fri, 08 Dec 2000 15:08:31 -0600

 Sorry for a delayed summary, we have been testing several configurations
before posting a summary.

 Looks like a lot of folks have this same question about Veritas Netbackup
support on TruCluster. Here is what a few had to say;

1) Several replies about Legato being the only "supported" solution with
TruCluster

2) Reports that version 3.4 will officially support Tru64 Unix 5.x, but not
TruCluster

3) And finally, Jim Jones reported that he has installed the agent on a
TruCluster 5.x using the cluster alias. Not supported, but appeared to be
working.

 So, I am talking to our Veritas Netbackup admin to see when we might
upgrade to 3.4. While we wait, I have installed the agent on the cluster
using the alias as suggested by Jim. This means that we will only run 1
instance of the agent (and not harness the power of all the cluster members
to shorten the backup window), but the agent will be available (via the
cluster alias) even if a cluster member is down. Seems to be working so far.
I am testing restores with ACL's (both Tru64 and ASU acl's) to see if the
acl's are contained in the backup image.

 Now, how about a multi-instance cluster aware agent that can somehow
checkpoint the backup session and recover from a cluster member loss without
having to restart the backup session? And be a supported platform.

That's about all for now, thanks to all those that responded.

Dan Goetzman

BMC Software
R&D Support, Houston
Dan_Goetzman_at_bmc.com

-- original posting follows --

 Anyone running Veritas Netbackup ( we are currently running v3.2) and have
the backup agent installed on a TruCluster 5.1 system?

 Is this supported?

 I am new to TruCluster, so I was thinking that the options might be;

 1) Install the agent on one node and have it back up everything, just like
a plain old Tru64 system. Maybe even configure it as a caa application so
that the agent would fail over to another member if we lost the member it
was installed on?

 2) Load the agent on each individual member and have each instance backup
1/2 of the data farm. This would shrink my backup window. If I lost a member
for any length of time, I would have to reconfigure netbackup to have the
remaining node backup everything.

 If anyone has played with this and has some input, I would sure like to
hear about it.
Received on Fri Dec 08 2000 - 21:09:19 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:41 NZDT