Hello everybody!
I'm using SuSE 9.2 + Postfix-2.2.2 + VDA-Patch 2.2.2 + Postgresql
8.0.1 + Courier-Imap 4.0.2 + Courier-Authlib 0.55
Aprox: 150.000 virtual users.
Everything works perfect, it is very quick doing querys to the DB,
DNS, etc, but.. from the last week, it became really slow when
"queuing the messages"
for example... sending an email manually...(domain is replaced by X):
telnet 127.0.0.1 25
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
220 mail.xxxxxx.net ESMTP Postfix (2.2.2)
helo xxxxx.net
250 mail.xxxxx.net
mail from:
250 Ok
rcpt to:
250 Ok
data
354 End data with <CR><LF>.<CR><LF>
hello world!
## UNTIL THIS POINT EVERYTHING IS REALLY QUICK:
## The telnet is quick
## helo, mail from:, rcpt to:, are quick
## This means that the postfix has already done every check
## DB, HELO_CHECKS, SENDER_CHECKS, CLIENT_CHECK, DNS, RBL, ETC
## but... when we end with "."
## It takes between 10 to 200 seconds to get the message queued
.
250 Ok: queued as 279C02555D0
quit
221 Bye
Doing a debug, there is not errors. no time-out.. nothing.. I just
cann see, a big time diference between the last INPUT of the client
and the QUEUED.
I was checking the process and the memory of the machine and it looks
ok, I even moved the DB to another server just to let this machine
work just with postfix... but.. it is just the same thing.
with a Netstat I see the amount of established connections on port 25
and with 2 or 70 connections it is the same, no difference.
The master.cf is intact from package, and in the main.cf I just add
the RBL, CHECKS, VIRTUAL, QUOTAS, ETC, I mean I didn't modify any
delay value, or time out, etc.
Some clue? idea? suggestion of what is happening?
some postfix value related to the queue that may be the problem?
somebody that has had this problem?
( I have never had this problem with postfix 2.0 )
tks and happy weekend!
--
Jota Omella
Developers Corp.
http://planetas.net