The goal is to get off of the aging V880 and onto a cheaper
Linux platform. At the same time, we've had a couple high-profile
hardware failures in the past couple months and, since the hardware is cheap
(and the VCS license is "free" thanks to our site license), putting
in a cluster seems a way to add reliability for the cost of a $5000 x86 box.
I don't intend to rename the master server, I'll move the name
and, I thought, make it the HA name for the pair - thus getting around the infamous
"can't rename the master server" problem. That's why I was
thinking this would be easy to pull off.
I don't know if the company will spring for $15k worth of
services, though. Kinda doubles the cost of the whole project.
From: Bryan Bahnmiller
[mailto:bbahnmiller AT dtcc DOT com]
Sent: Tuesday, October 13, 2009 1:58 PM
To: Donaldson, Mark
Subject: Re: [Veritas-bu] Converting stand-alone Master to clustered
Master
Mark,
Ugh...
Don't envy that task. A few comments.
The
master server has to run on the virtual name (relocatable IP and dns name.)
This name and the names of the individual nodes are in the EMM db. If you kept
the name of the current master as your virtual name, it should be a simpler
task. Also, the agent is very intrusive. (Especially on Windows.) There are
agent options that can only be set via the bpclusterutil command. That tells me
that the agent is not all that, what, conforming?
At the
conference I went to, a Symantec employee was asked why changing the name of
the backup server is not supported. He said that the name of the NBU master
server is in 14 different places on Windows servers and 11 places on Unix
servers. And the names in the different places have to be changed in the
correct order. One mistake and nothing works.
I
sat in on some very interesting discussions about NBU HA. What are you trying
to accomplish? Most of the people I talked with agreed that the main thing HA
will give you is the capability of doing server maintenance with minimal
impact. Most hardware now is pretty reliable. When was the last time you had a
nic or hba fail? When the server fails over, all backups stop. After a
failover, you will still have to clear out tape drives, restart backups and so
on. The only currently allowed HA configuration is active/passive. So you do
have one server sitting there idling, waiting for the other server to fail
over.
Also, if you have the disk mirrored with VxVM, you'll have to see if that
is supported with the Linux agent for failover too.
Bryan
|
"Donaldson,
Mark" <Mark.Donaldson AT Staples DOT com>
Sent by:
veritas-bu-bounces AT mailman.eng.auburn DOT edu
10/13/2009
02:32 PM
|
|
To
|
<veritas-bu AT mailman.eng.auburn DOT edu>
|
|
cc
|
|
|
Subject
|
[Veritas-bu]
Converting stand-alone Master to clustered Master
|
|
Page 14 of the NB High Availability guide:
=========
NetBackup does not support the conversion of an existing non-failover
NetBackup server to a failover NetBackup server. Contact Symantec
Enterprise Technical Support.
=========
Huh?
Should be easy, I'd think. Why would I need Tech Support?
Has anyone done this?
BTW: also converting from Solaris to Linux at the same time if I can
pull this off...
-Mark
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
_____________________________________________________________
DTCC DISCLAIMER: This email and any files transmitted with it are confidential
and intended solely for the use of the individual or entity to whom they are
addressed. If you have received this email in error, please notify us
immediately and delete the email and any attachments from your system. The
recipient should check this email and any attachments for the presence of
viruses. The company accepts no liability for any damage caused by any virus
transmitted by this email.
