[Mimedefang] Tracking down high memory usage
David Nelson
dave at itsdave.tv
Sun Apr 30 18:04:35 EDT 2006
Does anyone have advice on how to track down high memory usage with MD?
I run between 10 and 25 slaves and they all consume about 25MB of memory
at startup and during mail processing. After a while, _some_ of the
slaves (anywhere from 1 to 4) jump to about 175MB of memory (snips from
top below).
-----------
last pid: 73436; load averages: 1.50, 1.51, 1.95
up 9+02:40:26 15:00:46
127 processes: 1 running, 126 sleeping
CPU states: 1.6% user, 0.0% nice, 7.6% system, 5.6% interrupt, 85.2%
idle
Mem: 298M Active, 687M Inact, 223M Wired, 71M Cache, 112M Buf, 729M Free
Swap: 2048M Total, 7856K Used, 2040M Free
PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU
COMMAND
73423 defang -8 0 23392K 22548K piperd 0 0:02 38.81% 8.59% perl
73410 defang -8 0 23392K 22548K piperd 0 0:02 16.84% 7.13% perl
73370 defang -8 0 23392K 22548K piperd 1 0:02 4.45% 3.17% perl
73360 defang -8 0 23392K 22548K piperd 1 0:02 3.76% 2.83% perl
-----------
last pid: 73655; load averages: 0.62, 1.21, 1.78
up 9+02:41:57 15:02:17
114 processes: 2 running, 112 sleeping
CPU states: 53.3% user, 0.0% nice, 1.2% system, 0.0% interrupt, 45.5%
idle
Mem: 408M Active, 686M Inact, 224M Wired, 71M Cache, 112M Buf, 619M Free
Swap: 2048M Total, 7856K Used, 2040M Free
PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU
COMMAND
73161 defang 107 0 176M 135M CPU1 0 0:12 32.42% 32.42% perl
73611 defang -8 0 23392K 22548K piperd 1 0:02 5.41% 3.61% perl
73180 defang -8 0 24328K 23336K piperd 1 0:03 0.59% 0.59% perl
73527 defang -8 0 23392K 22548K piperd 0 0:02 0.42% 0.39% perl
-----------
I've tried and tried to figure out where that much memory would be
allocated from, but I've been less than successful.
Any ideas are greatly appreciated!
Thanks!
-- Dave
More information about the MIMEDefang
mailing list