06-17-2010 05:48 AM
The error that I got in crash dump is:
luster alias subsystem enabled
MEMORY CHANNEL API - initializing
Environmental Monitoring Subsystem Configured.
trap: invalid memory read access from kernel mode
faulting virtual address: 0x0000000000000000
pc of faulting instruction: 0xfffffc000065d1f4
ra contents at time of fault: 0xfffffc0000de3fcc
sp contents at time of fault: 0xfffffe0544ec74c0
panic (cpu 1): kernel memory fault
syncing disks... dump device name: dsk21-01, num: 0x130015d, off: 0x10, len: 0x5dc000
LSM attempting to dump to device at major 19 minor 349
dump device name: dsk1b-01, num: 0x1300288, off: 0x0, len: 0x5dc000
LSM attempting to dump to device at major 19 minor 648
Any idea what cause the server to crash?
06-17-2010 06:56 AM
If the system was able to reboot after the crash, there should be a recent crash_data.x in /var/adm/crash/ then please post that here.
The posted output is not sufficient to determine a possible cause for the crash.
07-13-2010 11:35 PM
One (.9) is related to virtual memory handling and the other (.10)
does look like a issue with the aha_chim (SCSI-Adapter) - Driver.
I guess your next question is..."What can I do to avoid this ?"
There are two points that I would go for...
1-the crashed node is a AlphaServer DS25 running on Firmware: 6.8-18
the most recent Firmware is Version 7.3 ! See:
I would go for a update of the Firmware.
(Don't forget to check the other node for this too !)
2-if you have to stay on 5.1A, well...go and look if you have all
additional early release patches installed, if not get and install them.
Better would be a consideration if you can go/upgrade to 5.1B-5
but that is only my humble opinion.