@Erik Seeing a lot of memory pressure on the server. Any idea where it's coming from or better still what we can do about it? I changed the configuration to kill processes that put pressure on memory instead of letting it swap.
That has greatly helped with responsiveness and stability, but is making things like Jenkins CI testing fail on the regular (kill's even simple rt's for some reason during distcheck).
jenkins is sitting on over 2 gigs resident, mysql is #1 at 867 megs, then we're down into <=100 meg procs
top and o res :)
how much is apache and the other dbs at? Last I looked, it was loads too, just spread across lots of child processes
I think server has 16gb
(don't have a console up atm)
bz% ps axu | grep 'www.*http.*T$' | awk '{print $6}' | xargs | sed 's/ /+/g;s,.*,(&)/1024,' | bc -l 613.26171875000000000000
613 megs, like half of mysql and 1/4ish of jenkins
system seems to have plenty of available memory...
yet it still runs out on the regular. I think something spikes it up. Maybe a botnet attaching and sucking up memory on the mailer or httpd (it's happened before), or maybe just regular activity coinciding. hrm. :(
Last updated: Jan 09 2025 at 00:46 UTC