I'm running tests on a multi-box setup (1xLDAP, 1xSMTP, 1xMailstore) now to spread the load a little. I'm sending about an avg of 5 messages every 10 seconds, from 5k to 14mb across about 10 users.
All is fine for a while and mail is delivered as expected. But then it simply seems to be queued and held in the active postfix queue (currently sitting with ~1100 messages in it and growing all the time).
The load on the SMTP box is sitting at :
The load on the other two is very low. /var/log/zimbra.log on the smtp box is simply showing this for each new message :
12:00:58 up 1:39, 3 users, load average: 0.26, 0.42, 0.46
total used free shared buffers cached
Mem: 514484 499396 15088 0 23976 267316
-/+ buffers/cache: 208104 306380
Swap: 1048568 232 1048336
Running /opt/zimbra/postfix/sbin/postqueue -f doesn't seem to do anything. There have been no mail deliveries to any of the users for an hour at least.
Feb 17 12:02:45 p317-02 postfix/qmgr: 920C83788D6: from=<root@xxxxx>, size=5087, nrcpt=1 (queue active)
Feb 17 12:02:45 p317-02 sendmail: k1HC2jJm004090: to=user@xxxxx, ctladdr=root (0/0), delay=00:00:00, xdelay=00:00:00, mailer=relay, pri=34619, relay=[127.0.0.1] [127.0.0.1], dsn=2.0.0, stat=Sent (Ok: queued as 920C83788D6)
Feb 17 12:02:45 p317-02 postfix/smtpd: disconnect from localhost.localdomain[127.0.0.1]
All the boxes are regular CentOS 4.2, no VM's etc.
Edit: about 20 or so messages were just delivered to each user, but 1400 are now in the active queue....