Lorian Development wrote:
A question for all you ASSP experts (I'm not one yet unfortunately):
ASSP 1.3.3.1 is going unresponsive for periods of 1.5 to 4 minutes,
every 10 minutes or so.
Only noticed this after upgraded from 1.3.1 to 1.3.3 (maybe a
coincidence?)
ASSP Config file in use since v1.2.6 (couple of months).
When this happens, CPU usage jumps to 100% during the period, it's
usually 4% to 8%.
During that time, the ASSP GUI will not respond, and local outgoing
SMTP fails (through ASSP).
During that time, the POP and IMAP DOES respond (on same machine)
TOP shows PERL using most of the CPU cycles.
Have examined the logs following episodes, haven't noticed anything
unusual.
Running on Mac PPC-G4, OSX 10.3.9 - also running mail server.
Perl v5.8.1
Not running ClamAV.
Delaying/Greylisting is ON.
Handling about 3,000 email connections per day.
Is a problem because outgoing SMTP fails -- users complaining.
A poorly written regEx could be the culprit. There were a number of
regex handling changes in 1.3.3 for some of the features, possibly
this
might cause issues with one or more regex strings.
Bayesian analysis on a large messages with a heavily loaded system can
also cause these symptoms.
It could also be related to the running-assp-as-non-root performance
issue some reported seeing on Linux.
Or it could be an external service causing the delay.
What do your logs show just before the CPU usage increases?
Kevin
Kevin,
I'm using the same ASSP config file that I started with on 1.2.6.
Would the regEx have remained the same as it was in 1.2.6 then?
Are there "standard - official" regEx's we should be using with 1.3.3
that are different that the ones supplied with 1.2.6?
I have not manually changed any of the regEx, except adding the line:
you\'ve.received.a.(postcard|greeting|ecard).from.a.(colleague|
your.partner|family.member|neighbor|mate|school-mate|school.friend|
worshipper|friend|partner)
(colleague|your.partner|family.member|neighbor|mate|school-mate|
school.friend|worshipper|friend|partner).has.sent.you.(a|an).
(postcard|greeting|ecard)
to the bombre.txt. That was provided by someone on this list a couple
of weeks ago,
to help catch all those recent "greeting card" spammers.
ASSP is running root/wheel - as a daemon - shouldn't be any problem
there.
Only "external" services running are IMAP and POP and they respond OK.
TOP shows Perl using all the CPU cycles during these episodes.
Bayesian analysis shouldn't have any different load than it had two
weeks ago.
Haven't found anything unusual in the logs - except these:
Aug-6-07 15:15:54 Cache reset (/usr/local/assp/spamdb)
Aug-6-07 15:16:13 Cache reset (/usr/local/assp/spamdb)
which seem to occur around the time of the episodes. I don't know if
they
are normal or not.
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Assp-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/assp-user
