We have a weird problem which I hope someone can help me with.
We have a 15 node cluster of HP servers running NW6.5 SP3 servers.
Each of the servers in the cluster are connected via two Broadcom NICs
to separate Cisco routers/switches and we are using the latest version of
HPs QASP teaming driver to provide load sharing and redundancy
on the two connections.

Now the problem. Sometimes when we move a resource from on
server to the other, users from outside of the server subnet cannot
access the resource. A network trace shows that the new
server properly sends an ARP to tell the network that there is
a new MAC address for the IP, but a little while later we seem
to see that the old server also sends an ARP and the
routers re-learn the old address. On the other hand, we have
no problem in accessing the resource from a client within the
server subnet - which is why I say that it "seems" that the
old server is sending an ARP.

There were ARP problems with the old version of QASP, but these
were supposedly fixed in the latest 2.25 release. Has anyone
seen a problem like this? Any ideas if this is a NetWare problem,
an HP problem or a Cisco problem?