LPr in a two-node cluster dies off frequently.

[Benny Koh]

Hi,
I have a serious problem with one of my LPrs in a cluster.

Hardware: a)2 x PIII 500Mhz LPrs (LPR-A and LPR-B), 2 x 9GB SCSI each, 256KB SDRAM each.
b) RS/12 with 12 x 9 SCSI, 3 RAID1 arrays (total 6 disks) and 2 RAID5 arrays (total 6 disks)

Software: a) Windows 2000 Advanced Server (SP3) b) SQL Server 2000 Enterprise Edition (SP3).
c) Microsoft Cluster Service (MSCS).

Cluster configuration: 2 SQL Server instances, one on each node.

Problem: Since 3 months ago at least, LPR-A has been dying off at least once a day (I don't know how else to call it). The monitor screen goes blank, cannot be pinged and cannot be accessed. The SQL instance on LPR-A will fail over to LPR-B as a result. However, the server fan and power LEDs are still lighted. When I manually restart LPR-A, a "the previous shutdown at was unexpected" appears in the event viewer. There is nothing in the SQL logs or event viewer that suggests anything out of the ordinary.

I have tried the following:
a) used a hardware diagnostic application to check the server, nothing wrong found.
b) swapped to a new harddisk.
c) installed only the base OS without cluster service.
d) reinstalled all applications.
e) removed LPR-A from cluster and stopped the cluster service.
f) upgraded the BIOS.

After doing all these, the problem was still there. I have pretty much at my wit's end. Can anyone help me with this?