280R Server keeps crashing or freezing

[yowza]

I have a Sun 280R running Solaris 5.8 that has just started to crash or freeze about every other day or so. I can't find anything in the /var/adm/messages file. I ran the prtdiag and everything is fine. I have not added any patches or touched this system in several months. I suspect it is probably hardware related, however, I can't find anything to prove this. I have it on a KVM switch and cannot access it, and can't access it through the serial port when this happens. All I can do is power it down hard and power on again to get it back up. Anybody have any ideas on how I can track this down?

Thanks,
yowza

 

[cndcadams]

Couple things to thinks about.

Memory or Software generally will cause hard hangs. So when the system hangs you can not do a stop A?
Have there been any additions to the KVM switch. I have NTI boxes which I have had issues on.(I will expand on this if you have added new devices to KVM switch).

Does the system actually ever reboot or does it just hang or crash? Does it power off?
If the system has rebooted it may have created a vmcore.* and a unix.* file under /var/crash/"hostname". If you have these files let me know.

Thanks

Chris Adams

 

[mikeclark]

A system hanging with no respone at the console can also be saused by a faulty device hanging the scsi bus. Check if your messages has any warnings about scsi devices.

 

[yowza]

Thanks for the responses Chris and mikeclark. When the system hangs I cannot access it via the console or serial port using a laptop. The keyboard does not work either so I can't do the Stop A. There are no messages in the messages file and no core dumps. I took it off the KVM switch last night and finally got it to boot. I set the diag-level to max but could not get it to boot in diagnostic mode. I tried running the obdiag but that wouldn't run either. I got a "Rejecting Alloc Mem" message. It hasn't crashed again yet, but it hasn't been up for 24 hours either)

yowza