What cause my RHL7.3 server to reboot?

[JDL2]

All,

I found that my server rebooted and the messages and boot log files do not indicated why it rebooted. All it contains is login information and all of a sudden syslogd startup. My boot.log file only has information starting from reboot and the boot.log.1 and boot.log.2 are empty. The other boot.log.* files are too old. Is there another way to find out why it rebooted?

Thanks,
JDL2

 

[franklin97355]

Have you ruled out power loss?

 

[JDL2]

Hi franklin97355,

No power loss here. Got a huge generator that kicks in if we do have a power loss and was just tested last week. I do have 30 or so servers that did not boot. Yes, power loss sure wasn't it.

JDL2

 

[unixrocks]

I remember having a problem very simmiliar to this except with a Sun Machine. The reson it was rebooting itself was due to an unseated fan on the memory. Could it be that the machine is overheating somewhere, perhaps a dodgy fan?

 

[unixrocks]

Forgot to add there was nothing in the /var/adm/messages file as to why it was rebooting.

 

[fluid11]

unixrocks is correct, problems like these are usually hardware problems, or overheating issues. What kind of server is this? If its a real server, and not one you put together yourself, I would contact the vendor and they will probably have you run some diagnostic tests on the hardware in the machine. If this is something you put together yourself, open the case up and look for the obvious things like the heatsink not being properly seated on the CPU and stuff like that.

ChrisP
RHCE, LPIC-1, CCNA, CNE, MCSE, +10 others

 

[JDL2]

Is there a utility that I could use to point me to what hardware may be failing? I checked connectivity on the lan cables and power cables. Unfortunately this server was purchased and no warranty was purchased with it. Its one of those servers that are sold by a VAR and support on them are really not that good, so I figured whoever purchased them for us before I got here thought it would be a waste to pay extra that would not be helpful. In HP, they have STM or EMS that would actually send alarms if something has failed. I'm hoping that RHL would have something similar. Any ideas? Thanks again. JDL2.