CPU overloaded
-
I'm running release 1.0 on a 1.1GHz Celeron with 128MB RAM (it's an old Dell PowerEdge server from a few years back - it even ran Exchange Server at one point!). I keep finding, though, that the CPU is at 100%. Running 'top' shows me that the top two processes are:
miniupnpd 67%
syslogd 28%So between them, those two processes are munching 95% CPU. I'm not sure that either is doing a great deal (syslogd doesn't seem to be picking up logs from any of my installed packages). There are usually about 8-12 users behind the PFSense box on my LAN, plus up to about half a dozen VPNing in. Is this kind of CPU usage normal, or is there something amiss here? Is a 1.1GHz Celeron a bit on the low end? (Screen dump from 'top' is pasted below). Could it be the CPU usage that's causing my occasional dropped IMAP connections?
Also (unrelated): the blog post announcing version 1.0 doesn't seem to have/link to a list of changes since the last RC/beta. Does anyone know where I can find that (apart from the CVS)?
last pid: 82533; load averages: 2.57, 2.47, 2.48 up 6+00:44:17 11:34:46
37 processes: 2 running, 35 sleeping
CPU states: 54.1% user, 0.0% nice, 41.7% system, 4.1% interrupt, 0.0% idle
Mem: 61M Active, 9276K Inact, 32M Wired, 4696K Cache, 22M Buf, 10M Free
Swap: 256M Total, 54M Used, 202M Free, 21% InusePID USERNAME THR PRI NICE SIZE RES STATE TIME WCPU COMMAND
719 root 1 132 0 3536K 1048K RUN 93.2H 68.80% miniupnpd
188 root 1 122 0 1360K 352K select 48.2H 26.03% syslogd
30398 root 4 20 0 52492K 8032K kserel 13:35 0.00% ntop
469 root 1 8 20 2296K 1004K wait 3:31 0.00% sh
30542 nobody 1 96 0 3928K 3292K select 2:02 0.00% darkstat
725 root 1 -58 0 23764K 20408K bpf 1:45 0.00% snort
334 root 1 4 0 3372K 1276K kqread 0:56 0.00% lighttpd
727 root 1 8 20 1192K 228K nanslp 0:55 0.00% check_reload_st
263 root 1 -58 0 3800K 1220K bpf 0:19 0.00% tcpdump
714 _ntp 1 96 0 1260K 372K select 0:12 0.00% ntpd
69245 root 1 4 0 42340K 13872K accept 0:11 0.00% php
434 proxy 1 4 0 656K 252K kqread 0:04 0.00% pftpx
30543 nobody 1 4 0 1516K 584K sbwait 0:03 0.00% darkstat
531 root 1 8 0 1304K 524K nanslp 0:03 0.00% cron
81735 root 1 4 0 42256K 13740K accept 0:02 0.00% php
264 root 1 -8 0 1196K 440K piperd 0:01 0.00% logger
81739 root 1 96 0 5568K 2236K select 0:00 0.00% sshd -
719 root 1 132 0 3536K 1048K RUN 93.2H 68.80% miniupnpd
Your hitting the miniupnpd race bug. Restart the service from Status -> Services when this happens. Check over on the miniupnpd thread for more info.
-
Thanks for answering. I couldn't find the miniupnpd thread (neither forum nor google search had much to contribute). Could you please point me in the right direction? For the record, my syslog is chucking out messages like this:
Oct 27 09:07:25 miniupnpd[93719]: Unsupported HTTP Command SUBSCRIBE
Oct 27 09:07:25 miniupnpd[93719]: Unknown soap method
Oct 27 09:07:25 last message repeated 903899 times
Oct 27 09:04:09 last message repeated 2762463 times
Oct 27 08:54:07 last message repeated 2764000 times
Oct 27 08:44:07 last message repeated 2760011 times
Oct 27 08:34:05 last message repeated 2783299 times
Oct 27 08:24:05 last message repeated 563501 times
Oct 27 08:22:04 last message repeated 139653 times
Oct 27 08:21:33 miniupnpd[93719]: Unsupported HTTP Command SUBSCRIBE
Oct 27 08:21:33 miniupnpd[93719]: Unknown soap method -
http://forum.pfsense.org/index.php/topic,551.0.html
-
Please reinstall the package. Those errors have been fixed. The latest version is 20061023.