Chris,
I read the info on Chapter 6 regarding Backup Managers, but I have a few
questions that I didn't see addressed in the manual?
1) What issues arrise when another manager node takes over the "containers"
with the IP address?  The current IP will no longer be the "center of the
universe" on the backup machine, how do we address that?  and what if the
backup manager has containers w/ the same name? can this be done?
2)  Can you have the backup manager node manage the the entire network?  In
the manual it sounds that you should break your network up into different
containers and assign the containers to different backup managers?
3) Does the Backup manager have to  be on a same release as PROD?  Will it
still take over?


> -----Original Message-----
> From: Chris Cowan [SMTP:chris.cowan AT 2ND-WAVE DOT COM]
> Sent: Thursday, February 17, 2000 2:16 PM
> To:   NV-L AT UCSBVM.UCSB DOT EDU
> Subject:      Re: Implementing Netview on HA
>
> "Ferreira, Linda" wrote:
> >
> > Chris,
> > I will look into Chapter 6 on Backup Managers.  Unfortunately, I don't
> know
> > too much about HA myself, but we are an HA shop and we have some
> "experts"
> > which have implemented HA on other applications but unfortunaltely they
> > don't know Netview.  Their question for me was "Is  Netview HA'able?"
> > meaning, could it reside somewhere other than rootvg, if so, it could be
> > done.  All they need to know is whether or not  Netview can be installed
> on
> > a seperate logical volume (not part of the operating system) Can you
> answer
> > that question for me being that you have implemented it for your
> customer?
> > Any assistance would be greatly appreciated.
> >
>
> Linda,
>
> There is no problem placing or moving the /usr/OV filesystem to any
> volume group.   The real trick is handling the Netview databases.
> Netview builds the entire topology DB with itself as the point of
> reference.   Depending on how you do the NIC configurations on the HA
> machines, you may or may not have to reconcile the databases (which
> could involve running reset_ci, ovtopofix, etc....) when you failover.
> If you have to do this, this could increase the failover time
> dramatically.
>
> Also, if you're going to put /usr/OV in a shared disk array, you're
> going to also have to make sure that you kill all processes (like the NV
> webserver) using files in /usr/OV.  You may also have to recycle snmpd,
> in order to make sure that you have the proper SNMP stack configuration
> underneath of Netview.
>
> I recommended the Backup Manager approach because it's just a lot
> cleaner, both from an installation/config and maintenance standpoint. <<
> File: Card for Chris Cowan >>