Today, I noticed a strange dump summary for two disk partitions. This
was mysterious to me because I'd been running without problems for
about 10 nights in a row. A brief glance at the messages e-mailed to
operator was enought to tell me, that all the problems were caused by
me being logged in from home at the time. This is a small public
library and we are very closed at night. All the problems were caused
because of open sockets and one or two open files.
Does anyone have an example wrapper script that will shut down
processes and then restart them after their relevant files ,mostly log
files in my case, have been backed up. Since we are closed at night,
the simplest thing to do would be for me to shut down my proxy server
(squid - and no I don't back up my cache), web server (apache), and
SQL server (mySQL), while the affected partitions (/ and /var )are
being backed up. I am curious if anyone has done this. Ideally, I
would like to stop and start the processes in the middle of the run
and have something detect when it was OK to start them up
again. Though, since we aren't anywhere near a 24/7 operation I have a
lot of leeway.
--
Josh Kuperman
[EMAIL PROTECTED]
