System is an old Compaq Presario, Athlon XP 2000+, 512M RAM (~504M after 8MB shared with graphics). 1GB swap, kernel 2.6.32-642.3.1.el6.i686
Yeah yeah, add more RAM, buy a real server, etc... but even shiny new can run into stuff, trying to learn how to troubleshoot this.. am stuck
Really only running httpd, mysqld. Has a small WordPress site that isn't accessed very much, an rysnc task that runs each hour to snapshot the drive and runs an fs dump (on LVM snapshot) each night... assuming it gets to that point.
Was running OK when first installed a few months ago, even with using DropBox CLI to keep backup files in sync on Dropbox... Have stopped dropbox service trying to narrow down issue.
Now past several weeks, I am getting hard crash - system can't do anything without a hard reboot, /var/log/messages stops short of showing the properly - like this last time the last message is "dhclient: No working leases in persistent database - sleeping"
Before this are several OOM-killer runs at about ~480M used ( +/- buffers, even... ), killing mysqld and httpd tasks first, then ntpd, etc... until nothing can be killed...
I added a cron job to logger output from free -m every 5 minutes - once OOM starts, I don't see those messages in the logs, but there is steady creep from ~200M used at boot to this ~480M where the system freaks out. Swap only gets about ~10M used at most.
So if nothing is running on the system at that point, what is causing the crashes?