Performance with postfix
glenn.steen at gmail.com
Thu Apr 2 16:12:45 IST 2009
2009/4/2 JC Putter <jcputter at centreweb.co.za>:
> Hi i am busy testing what load my mailscanner box can handle.
> i am running an intel 2.4Ghz with 1GB RAM and a normal SATA HDD drive using
> Postfix as MTA all RBL's are in SA al running on Centos 5.2
> i am using smtp-source to test sending 2000 messages at once and i not
> getting the results i was hoping for it, after 24 min of running there are
> still 1300 messages in the queue.
> is the normal or bad, postfix is a fast mta, i also noticed that mailscanner
> scans messages in batches of 30, and it really takes a long time for the 30
> messages to be placed back in the queue.
> procs -----------memory---------- ---swap-- -----io---- --system--
> r b swpd free buff cache si so bi bo in cs us sy id
> wa st
> 0 0 187204 43488 19036 65140 1 1 7 23 73 7 5 0 94
> 0 0
First line of vmstat is pure drivel, run it like "vmstat 2" so that
it'll poll every 2 seconds... Pay attention to line 2 and onward
(during a "stressful situation", of course:-). The goal is to use as
much RAM as possible without swapping (then you are "rightsized:-).
> again am i asking too much from the hardware or is there something i can do
> to increase performance? '
I think you are low on RAM (at least), and I'm guessing you might see
some troubles with network lookups too, although you'd need
corroborate that (do some trial runs with MailScanner --debug
--debug-sa and see where the time is really spent).
> i am using a caching dns server and added
> none /var/spool/MailScanner/incoming tmpfs defaults 0 0
> in fstab
Which eats at your RAM too... Try reducing the amount of MS children
and see if that actually helps...
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se
More information about the MailScanner