2-Node Cluster questions
Novell Cool Solutions: Question & Answer
Reader Rating 
Q:
Michael W. wrote: I have have just completed a NW 6.5 Sp3 2-Node Cluster and iSCSI Target deployment Novell Cool Solutions Appnote Mar 23 2004. Each cluster node houses two ethernet adapters and supports a public and private IP network- respectively.
On the public network the cluster nodes 1 and 2 are 192.168.123.121/24 and 192.168.123.122/24. On the private ( heartbeat) network the cluster node 1 and 2 are 10.0.0.121/8 and 10.0.0.122/8. The ISCI Target server (RAID5)IP address is 10.0.0.120. The share cluster volume is name is "Sandbox" and resides on cluster-node 1.
A:
Q1: When I down cluster-node 1, the shared volume sandbox fails over as expected to cluster-node 2. When I reboot cluster-node 1, the shared volume does not fail back. Any idea why?
A1: Auto failback is disabled by default. This can be changed pre cluster resource if auto failback is desired.
Q2: If cluster-node 1 (Pub/Net 192.168.123.121/24) is downed while I am attached to and working in the shared volume "Sandbox," access is retained. However if I reboot the client and try to login into cluster-node 1 again the login fails. How do I configure the cluster-nodes to fail-over each others' IP address?
A2: Does the client depend on cluster-node 1 in order to login to the tree?
Try instead logging into the tree via the cluster virtual server's IP address.
Q3: Am I hurting myself by running public and private network?
A3: We recommend having cluster heartbeats on the same LAN as client/server packets.
Novell Cool Solutions (corporate web communities) are produced by WebWise Solutions. www.webwiseone.com