We have 2-node cluste with ISCSI node, it runs on nw65sp5, each server
contains two nic cards, with Broadcom NIC connected to the network LAN on
cisco 3750 stackable switch, and Intel e1000.lan connected to the ISCSI
box using an unmanaged 8 port gigabit dlink switch

We have recently update our two node-cluster to the latest patches as below
driver for b57.lan to version 9.24
driver fro intel pro/1000 pci/pci-x dated on 1 july 2005 version 8.14

After the those updated above, one of your cluster node, which node2
intermittently get disconnected from the network (I means the network
connected to the cisco switches). It causes the ip address conflict, as It
can't see the network and try to do a failover as it thinks the node1 has
failed. And I have to perform the uldncs and ldncs to reload the cluster
service to join back to the cluster for node2

Any idea