Hi, On Fri, Mar 22, 2002 at 06:21:12PM +0800, Julian Gomez wrote: > On Thu, 21 Mar 2002, Stephen C. Tweedie wrote: > > > That could be anything --- hardware or software. Are there any kernel > > logs when this happens? > > Unfortunately no. Log file just stopped at 14:34 and restarted back at > 15:50 when it was hard-resetted. You should probably set up a serial console, then. See http://people.redhat.com/johnsonm/ttywatch.html for software to let another host act as a monitor and log collector at the other end of the serial console line, and the Linux source code in file linux/Documentation/serial-console.txt for instructions about setting up the kernel to copy log messages to a serial port. I use this on all my test boxes and it makes tracking down problems like this much easier. > > Do you still have kernel console access? > > Yes. SSH 24/7 kernel console != network access. > > Can you trap any oops messages that might occur in this state, or can > > you trigger altgr+sysrq+t or altgr+scrlck kernel debugging dumps? > > I'll have them try it the next time. Hopefully something comes up - alas I > can't do it (remote from one branch - 20 miles; another is about 250+ > miles). Understood. Next time you get to one of the boxes, I really do recommend setting up the serial console if you can. The ttywatch program lets you set up a telnet port as a relay on the logging host so that you can then telnet in to see the kernel log messages, as well as logging to /var/log/ttywatch/*. Cheers, Stephen
