Hello Ofer,
I say "very idle" because I have 30/70% cpu idle and 10/30% cpu wait.
Sorry I checked...
- Get a faster disk
I have 2x36Go U320 15000rpm mirrored with md, 2x73Go U320 15000rpm mirrored
with md, and 4g0 memory (cache).
- Put your postfix queues on a separate disk from other stuff
I have on the 36Go, the system, the postfix/puremessage code, the postfix
logs/queues, on the 73Go the puremessage postgress database/quarantime/logs.
The top filesystems I/O are postgress and postfix logs, I have writes on the
postfix queues and few reads (cache disk).
- Mount that partition with the "noatime" option
I have this
/dev/mapper/vgroot-lvpmx on /opt/pmx type ext3 (rw)
# postfix/puremessage code
/dev/mapper/vgroot-lvpmxlog on /opt/pmx/var/log type ext3 (rw)
# puremessage logs
/dev/mapper/vgroot-lvpmxqdir on /opt/pmx/var/qdir type ext3 (rw)
# puremessage quarantime
/dev/mapper/vgroot-lvpfxlog on /opt/pmx/postfix/var/log type ext3 (rw,noatime)
# postfix logs
/dev/mapper/vgroot-lvpfxspool on /opt/pmx/postfix/var/spool type ext3
(rw,noatime) # postfix queues
/dev/mapper/vgroot-lvpgsdata on /opt/pmx/postgres/var/data type ext3 (rw)
# puremessage postgres database
- Put your postfix queues on tmpfs
I can not lose mails...
Best regards.
Francis
________________________________________
De : [EMAIL PROTECTED] [EMAIL PROTECTED] de la part de Ofer Inbar [EMAIL
PROTECTED]
Date d'envoi : mardi 28 octobre 2008 17:50
À : postfix-users@postfix.org
Objet : Re: Big incoming queue, slow qmgr, idle system.
Francis SOUYRI <[EMAIL PROTECTED]> wrote:
> Time to time we have lot of mail in the incoming queue, "no" mail in the
> active queue, the "qmgr "is very" slow and the server is very idle. We
> tested the antispam/virus to see if there is a problem but the response
> is very good.
When you say "very idle" what do you mean? Have you looked at disk,
or just CPU? If you haven't looked at disk as a possible cause of
your bottleneck, try running "vmstat 1" next time you see this
happening and look at the last two columns, id and wa (idle cpu, wait
cpu). If both are high, then qmgr may be slowed down by I/O wait not
CPU.
You may already know this, and you may already have checked for this,
but I think if you had you'd likely have mentioned it in the email, so
since you didn't mention it, maybe you didn't look at it.
Possible solutions:
- Get a faster disk
- Put your postfix queues on a separate disk from other stuff
- Mount that partition with the "noatime" option
- Put your postfix queues on tmpfs
(but you could lose mail in a crash)
-- Cos
