Hi All,

I have just setup a cluster and am testing it's strengths & weakness's in
our environment.
If I shutdown the active server, the passive server does it's job and
becomes the active server (whether I gracefully shutdown the first server,
or pull the power lead out)
When I simulate hard disk failure by un-seating the disks in one of the
SAN's, data & storage is still available. After reseating the SAN disks,
data is automatically re-mirrored and all appears ok (only tested under
minimal load so far).

Here is my problem:
Assuming 'server A' is active & 'server B' is passive at the start of this
When I unplug the network cable from the Nic (simulating port failure of the
switch), I had hoped that 'server B' would then become the active server BUT
instead it gets a Poison Pill (Abend on P00: Ate Poison Pill in
SbdWriteNodeTick given by some other node)...

Is there some config or fix to egt around this limitation? As it seems to
defeat the point of High Availability,,,
(I realise the confusion arises from both fibre links still being active)

Two node Netware6.5 sp5 Cluster
Mirrored storage hosted by two Fibre connected SAN's (each attached to each
respective cluster node)