Re: how to determine why system rebooted
I thought this would be easy to discover but I've just come back and my
server has rebooted itself. From various /var/log files I can see when
but I can find NO reason or even hints why.
I've poked Google and peeked the Debian web pages but have come no
closer to this.
To risk being redundant, a power outage will leave no clues, and the only
way to know that happened is a smart UPS with logging that you can check
after the fact.
The next most likely cause, I think, would be a system crash resulting
from unstable hardware or system software. In most cases this will result
in a kernel "oops" which is printed to the system console but for a variety
of reasons may not make it out to a system log file, and don't forget, even
if the output data doesn't make it to log file there are ways in linux to
cause a crash dump, which can then be sifted through using the gnu debugger
to often locate the exact cause of the crash. (Sorry, I'm rusty on the details.)
If you are serious about pursuing the last possibility, then there is
a kernel build option that sends sends the system console output out
through some I/O devices (such as parallel or serial ports, I don't
recall offhand). I even recall an option to send console output over
a network as "raw packet" data, or something similar. Perhaps one of
these would be helpful in your case.