> Maybe these links might help:
> 'Novell KMP: KMP’ing IBM’s RDAC driver'
> (http://blog.barfoo.org/2009/05/08/no...s-rdac-driver/)
> http://forums.novell.com/novell-prod...el-update.html
> 'Creating a Kernel Module Source RPM - Developer Community'
> (http://developer.novell.com/wiki/ind...ule_Source_RPM)

I have taken a look into this docs, but I don't think that this is the
way I should go. If something goes wrong I would be lost totally.

I'm not sure what will happen during the upgrade to SP2a. SLES SP3 will
bring it's own version of the HBA driver (qle2460) and I am afraid that
on reboot this driver will load. RDAC will not load (because of the new
initrd file) and therefor the server will see all SAN LUNs twice.

What will NSS / NCS do with that? Try to access these LUNs? Even try to
mount these LUNs?

I know that normally NCS will not mount a cluster volume that is in use
already, but under these circumstances does this rule still apply?

Best case would be that NSS does nothing (sees the LUNs but do not try
to access them) and NCS finds out that something is wrong (maybe by
seeing its cluster partition twice) and stops loading. But is it working
this way?

Can somebody can bring some light into this? :-)

Thanks again,