[INUG-Users] redundancy and availability on Netcool Omnibus v7.1

[James Farrow]

Ives that document is available on the GAT site I think or you can ask Don
Wildman (Product Manager) for it directly. His email address is on this list
in some recent emails.

There is also a property now available in 7.2 BI and UNI directional
gateways (think maybe in 7.1 too) which is called "Gate.Resync.LockType ".
This allows you to use FULL, PARTIAL or NONE as values for "locking"
ObjectServers (reader and writer) when resyncing.

Unfortunately this property is not documented as yet, at least I couldn't
find it in the docs. But if you ask IBM I am sure they will provide more
information on how it works. I think this could help you with your problem,
whilst you maintain the considerable benefits of the standard Failover
architecture.

Hope this helps

James





On 20/05/2008, ives.dekoninck at bt.com <ives.dekoninck at bt.com> wrote:
>
> Dears,
>
> I have a design and configuration question related to redundancy and
> availability on Netcool Omnibus v7.1.
>
> We have a setup more or less as shown below. We have configured a
> primary and secondary OS on each layer and configured the servers to
> connect to a primary server via a unigate if possible and to a backup if
> the primary is unavailable. Between the OS's runs a bigate keeping them
> in sync. The resync mode is NORMAL. Like it was in the previous 3.6
> version.
>
> At the moment when a failure occurs redundancy kicks in and all is fine.
>
>
> The problem starts when for instance the primary consolidation server
> comes back online after a failure. Before it has had time to synchronize
> all other object server want to connect to it and push in or collect
> their events. This causes the Netcool service to be unavailable until
> all synchronization is done and with the short SLA's as we have them
> this is not desirable.
>
> Since in all cases the hardware of the primary and secondary OS is equal
> my first idea is to forget about primary and secondary and always remain
> on whatever OS is available and remain there until it becomes
> unavailable. So basically no fallback to a primary server unless the
> secondary server fails.
>
> Many of the example configurations I have seen so far seem to
> religiously stick to a primary and backup schema where the gateways
> instantly reconnect to a primary when it becomes available
>
> I think some of you may have experience with a situation like this and
> can advise on a best practice. This is the advise I seek.
>
> A previous thread I found mentions a document on support.micromuse.com
> called  "Event Service Framework baseline architecture" but I was unable
> to find this on the ibm site. It may have been renamed. If anyone has it
> or knows what it is called I'd appreciate if you tell me.
>
> Best regards,
>
> Ives
>
>
>
> <mailto:users at netcoolusers.org>
>
> _______________________________________________
> Sent by the netcoolusers.org "users" mailing list
> Post: users at netcoolusers.org
> Unsubscribe: users-unsubscribe at netcoolusers.org
> Search: http://netcoolusers.org/Search
>
>