João Reis wrote:
I have a suse 9.0 professional server which is used as an development server. This server is 24 hours up and once and a while it simply crashes. The users cannot access it throught telnet and in the console the keyboard does not respond, so i must do a hard reboot (reset bottom).
I have checked and these crashes occur normally between 6am and 8am. I tried to see what was happenning at that time in the logs but the logs stopped responding at the crash time. So i cannot see the status of my machine at the crash time.
Fortunatly these crashes occur very rare. I can say that between September and December they have occured 2 times, but i wished i could know more why these crashes occur.
Is it a hardware problem ? How can i figure out ? If my logs does not show anything, how can i fetch more information about the system status at that time ? Is there any well known fix i should know abaut ?
I had a box that was doing intermittent crashes with no log entries, though not at specific times of day. It turned out to be faulty memory. Reboot the machine and run memtest86 (it should be listed as 'Memory Test' on the bottom of the GRUB menu). If that doesn't find a problem, another thing you can try is to connect some device (another computer?) to a serial port and direct all kernel messages to it. That way you may see something that is otherwise getting lost in the crash. HTH, Dave