Proliant 5500 reboots randomly with Linux

[Brian_317]

Hello group,

The system has been randomly rebooting for months now, and it seems to be more often.

The system is a dual Xeon 500 with 512mb ram, 3x9.1 gb hd attached to the internal scsi card. two drives are in a raid 1 software array.

OS is Fedora Core 1, though I have tried FC1 SMP and UNI processor kernels, as well as RHEL3 SMP and UNI processor kernels. Results are exactly the same.

I have verified the latest BIOS has been installed (see http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=605012), and I also have installed the IML viewer (cpqimlview) along with the management utilities allowing logging to syslog through the /proc/cpq??? system.

Temperature seems okay, and I assume this systems only has the single sensor:
]# cat /proc/cpqtemp ID TYPE LOCATION STATUS CURRENT THRESHOLD
1 LM75 Sensor System Board Nominal 73F/ 23C 138F/ 59C
2 LM75 Sensor System Board Nominal ---F/---C 147F/ 64C
3 TC620 Sensor I/O Zone Nominal ---F/---C 134F/ 57C

Fans seem okay, with only two installed. Not sure if RPM's should be listed here for this system:
# cat /proc/cpqfan
ID TYPE LOCATION STATUS REDUNDANT FAN SPEED
1 Basic Fan Processor Zone Absent No -------
2 Basic Fan Processor Zone Nominal No -------
3 Basic Fan I/O Zone Absent No -------
4 Basic Fan I/O Zone Nominal No -------

The BIOS has been setup to use "LINUX" as the operating system. I would really like to know what the difference is _exactly_ between the many modes available.

The memory has been tested using the Compaq toos memory test, as well as the memtest program from the latest Knoppix cd. It checks out fine.

There is nothing in the IML or syslog preceeding the crash indicating something went wrong at all.

The server is not heavily loaded.

I have read the threads with related questions, and the next recourse seems to be change the motherboard. This server is out of warranty, and it will be less costly to simply replace the entire box if it comes to this.

Are there any suggestions as to what to attempt next?


Thanks,

Brian