Windows 2000 Adv cluster on MSA 1000. Strange problem

[yusong_1]

HP MSA 1000. Two cluster nodes. Only two cluster nodes are attached to this device.

The problem is that:
1. Sometimes, nodeB will hang. Windows 2000 Kernel mode works. Therefore, the failover doesn't happen to nodeA.

2. We have no way, but reset the computer.

3. After we reset the computer, nodeA tried to take the SAP group on the nodeB. During the failover stage, all the disks are not connected. In the cluster log, error 1167 is reported.
We have to reboot nodeA again.

I have asked MS and HP to check this problem. MS thought that it was a SAN configuration or hardware issue. HP checked the hardware. They said no problem. This issue has happened several times. I don't want to solve NodeB hang issue. Cluster should be able to provide fault-toleance abliity. How do you think this problem? Can we enable the log at the switch level to check? I suspect that the switch embeded in MSA 1000 is reset when this behavior occurs.

In the event log on nodeA, we can see the event 9, 11, and 15 when nodeA tried to take the group SAP.

Thanks,
Sam

[David Campbell_4]

Check you dependencies and make sure network name and storage sets are correct. Check you your perferred owner as well.. You need to check each resource don't just look at the groups...