Daryl C. W. O'Shea writes: > On 30/03/2008 5:08 PM, Justin Mason wrote: > > Daryl C. W. O'Shea writes: > >> Seems to be doing a whole lot of something, all the time... > >> > >> load averages: 1.00, 1.05, 1.13 > >> 05:30:27 > >> 57 processes: 55 sleeping, 2 on cpu > >> CPU states: 49.7% idle, 50.0% user, 0.3% kernel, 0.0% iowait, 0.0% swap > >> Memory: 2048M real, 80M free, 716M swap in use, 7742M swap free > >> > >> PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND > >> 17388 automc 1 0 15 460M 450M cpu/0 320:16 49.99% > >> hit-frequencies > > > > Is this always the same process? generating the "overlap" report > > on a big set of logs can take a lot of RAM... > > Not sure... 320 minutes seems to be a while though. I had just noticed > that hit-frequencies had been running for a long time both before and > again after the zone got bounced the other day. I'm probably just > noticing it running longer now that there's over 2GB of input.
Yeah, it runs almost continually -- overlap is a big job. I may fix it to only do overlap reports for the nightly and weekly sets, if it doesn't already... here's what it's currently doing: : jm 5...; ps -elf | grep 25671 0 O jm 29142 1429 0 50 20 ? 265 09:05:51 pts/3 0:00 grep 25671 0 O automc 25671 25611 50 99 35 ? 123858 08:01:53 ? 63:39 /usr/bin/perl -w ./hit-frequencies : jm 6...; sudo truss -fp 25671 25671: write(1, " . 0 0 0 0 0 0 . 7".., 5120) = 5120 25671: write(1, " X X X X ; 0 % o".., 5120) = 5120 25671: write(1, " h i t _ _ P H D\n ".., 5120) = 5120 25671: write(1, " a l s o h i t _ _".., 5120) = 5120 25671: write(1, " U T L O O K _ E X P R E".., 5120) = 5120 25671: write(1, " E N T\n o v e r l a".., 5120) = 5120 25671: write(1, " % o f _ _ D O S _ S".., 5120) = 5120 25671: write(1, " s p a m : 3 8 % ".., 5120) = 5120 25671: write(1, " h i t s a l s o h i".., 5120) = 5120 25671: write(1, " S\n o v e r l a p ".., 5120) = 5120 [CTRL-C] --j.
