On Thu, Sep 18, 2008 at 06:34:23PM -0500, Len Conrad wrote:
> >> Suggestions?
> >
> > Yes, you should be running much more than 2 instances of amavisd.
>
> was running 6, upped that to 10 about 1/2 ago. no change.
>
> sockstat -4 | egrep vscan
> vscan perl 36980 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36979 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36978 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36977 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36976 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36975 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36974 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36973 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36972 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36971 5 tcp4 127.0.0.1:10024 *:*
> vscan perl 36919 5 tcp4 127.0.0.1:10024 *:*
>
> ps aux | egrep -i vscan
> vscan 36919 0.0 6.5 71688 66624 ?? Ss 5:09PM 0:21.34 amavisd
> (master) (perl)
> vscan 36971 0.0 6.5 72320 66848 ?? I 5:10PM 0:00.10 amavisd
> (ch1-avail) (perl)
> vscan 36972 0.0 6.5 72320 66848 ?? I 5:10PM 0:00.09 amavisd
> (ch1-avail) (perl)
> vscan 36973 0.0 6.5 72320 66816 ?? I 5:10PM 0:00.04 amavisd
> (virgin child) (perl)
> vscan 36974 0.0 6.5 72320 66816 ?? I 5:10PM 0:00.04 amavisd
> (virgin child) (perl)
> vscan 36975 0.0 6.5 72320 66816 ?? I 5:10PM 0:00.04 amavisd
> (virgin child) (perl)
> vscan 36976 0.0 6.5 72320 66816 ?? I 5:10PM 0:00.04 amavisd
> (virgin child) (perl)
> vscan 36977 0.0 6.5 72320 66816 ?? I 5:10PM 0:00.04 amavisd
> (virgin child) (perl)
> vscan 36978 0.0 6.5 72320 66816 ?? I 5:10PM 0:00.04 amavisd
> (virgin child) (perl)
> vscan 36979 0.0 6.5 72320 66816 ?? I 5:10PM 0:00.04 amavisd
> (virgin child) (perl)
> vscan 36980 0.0 6.5 72320 66816 ?? I 5:10PM 0:00.04 amavisd
> (virgin child) (perl)
>
> This is not a busy machine at the amavis/spamd/clamav level, until about
> 17:00 hours:
This confirms that while you have 10 configured in amavisd, only 2 are
being used. Note that all of them after ther first 2 say "virgin"
indicating that postfix has never connected to them.
> egrep -ic '(passed clean|spam-tag|blocked spam)' /var/log/maillog
> 10043
>
>
> >Your current 2 instances virtually guarantee that mail will back up
> >unnecessarily while SA waits for responses from DNS queries and other
> >network tests.
>
> was not two, was 6, now 10
Sorry, I was confused by your mention of "2 instances of amavisd".
I believe I see what's going on now.
> 230 MB RMA inactive or free.
>
>
> > You need more RAM
>
> we'll shoot for 2 GB
>
> > Try 6 amavisd processes for a start (Scientific Wild Ass Guess), and
> >adjust up or down from there as necessary. If it starts swapping, drop
> >the number of processes, of course. Make sure that the Postfix
> >concurrency setting for injection into amavisd is set to match.
>
> local concurrency was 2, upped now to 10.
>
> local_destination_concurrency_limit = 10
> default_destination_concurrency_limit = 10
Ahhh, I think that's your problem. I think these concurrency limits
are not applying to amavisd injection. Check your Postfix master.cf -
it may have a specific process limit of 2 set for the amavisd injection
instance of smtp. Alternatively, check whether your main.cf has a
"vscan_concurrency_limit = 2" for vscan or whatever you've called the
special amavisd injection processes in master.cf.
-- Clifton
--
Clifton Royston -- [EMAIL PROTECTED] / [EMAIL PROTECTED]
President - I and I Computing * http://www.iandicomputing.com/
Custom programming, network design, systems and network consulting services
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
AMaViS-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/
