Wayne Doust wrote in message news:<646bmvc88g143npvs4c3p5mq7nasnprg8f@4ax.com>...
> I have been asked a question that I don't have a definitive answer
> for.
>
> [Please note: You will have to be reasonably familiar with SAN and NCS
> to answer this question]
>
> I have a client who wants to connect two Netware 6.x servers with a
> small SAN that has two service processors. The client wants to connect
> each of the servers via only ONE Fibre Channel card to separate SP's.
> (I know, this is not how it's supposed to be done).
>
> The vendor will configure the SAN to failover the SP's between file
> systems.
>
> My question. Is it possible to configure Netware to detect the
> failover? I think not, as the LUN's are detected on driver load except
> when a SCAN FOR NEW DEVICES is initiated, and even then the LUN's
> would have to be mounted. To work properly the following would have to
> happen:
>
> 1) NCS detect the LUN failure on a node and initiate a reconnection on
> the correct node. This would need to be preceed by a script to scan
> for the new lun and then mount the file system.
>
> 2) A hack may be for a CRON script to continually scan and mount file
> systems every five minutes or so.
>
> Either way, if it DID work I think it would only work for File
> Services and nothing else.
>
> Anyone have any experience or insights with this?
>
> Regards,
>
> Wayne
>
> -----
> This is a test of the emergency signature system.
> If this were a real test, you would be reading an amusing anecdote,
> personal philosophy, or corporate disclaimer. This is just a test.

I am assuming NW6.X with shared storage on the SAN.
The 2 NW server see the shared storage all the time, however only one
of the servers a storage pool active any given time.

The way I see it (simplisticly) you have 2 situations that can happen:
1. One of your servers fail, this will be detected by NCS either by
the LAN heartbeat trigger or by an missing SBD update to the SAN. This
is handled by NCS as designed and services+pools running on the
failing server will be started on the running server, after Abending
the failing server (just to be sure).
2. One of the controllers in the SAN may fail, this causing the SBD
partion not be updated at the set interval (½ heartbeat setting) again
resulting in NCS in
reacting by Abending the "offending" server i.e. the server not
connected to the SAN anymore. At this point pools will be activated on
the remaining server and all is well.

Kind Regards,

Bjørn