I'm running Mac OS X Server 10.3.9 on a PM G5, and lately my users are experiencing well below average send and receive times: 20 to 30 seconds for small, text-only email. Nothing really changed and the ammount of email that we get is about 3000 x day. A Barracuda Spam firewall is in place stopping 95% of the spam.
The slow service was there even before the Barracuda system was installed.
I read some topic in this forum but they didn't solve the problem.
The DNS service on the machine is off. and all the SMTP or IMAP & POP connections are using internal IPs.
Not a lot of info there to go on. Your mail logs will show how long an indivisual message takes to go through the system (between a connect to client and corresponding send to relay server) - this would show where any 'delay' is taking place. Can you post this so folk can compare it to their own?
I'm running Mac OS X Server 10.3.9 on a PM G5, and
lately my users are experiencing well below average
send and receive times: 20 to 30 seconds for small,
text-only email. Nothing really changed and the
ammount of email that we get is about 3000 x day. A
Barracuda Spam firewall is in place stopping 95% of
the spam.
So spam & virus scanning on the server is OFF?
The slow service was there even before the Barracuda
system was installed.
I read some topic in this forum but they didn't solve
the problem.
The DNS service on the machine is off.
So does the server have any dns to refer to for its own hostname/ip ?
Here's some sample of the log
Jan 30 10:37:25 www postfix/smtpd[29615]: EDD571556A5F: client=unknown[192.168.0.203], sasl_method=LOGIN, sasl_username=xxxx
Jan 30 10:37:26 www postfix/cleanup[29616]: EDD571556A5F: message-id=<C1E4D063.DCA9%xxx@xxxx.com>
Jan 30 10:37:26 www postfix/qmgr[414]: EDD571556A5F: from=<xxxx@xxxx.com>, size=4164, nrcpt=1 (queue active)
Jan 30 10:37:31 www postfix/smtpd[29615]: disconnect from unknown[192.168.0.203]
Jan 30 10:37:39 www postfix/smtp[29645]: EDD571556A5F: to=<xxxx@xxxxx.com>, relay=xxx.xxxxx.com[216.234.54.66], delay=14, status=sent (250 Ok: queued as 041F8BEFC29)
Jan 30 10:56:06 www postfix/qmgr[414]: 6A8131556C8E: from=<xxx@xxxxx.com>, size=718, nrcpt=1 (queue active)
Jan 30 10:56:10 www postfix/smtpd[29615]: disconnect from unknown[192.168.0.227]
Jan 30 10:56:27 www postfix/smtp[29973]: 6A8131556C8E: to=<xxxx@xxxxxx.com>, relay=mx1.balanced.postal.mail.dreamhost.com[208.97.132.51], delay=21, status=sent (250 Ok: queued as 4C1AECC6E2)
yes virus and spam are off on the server and running on the barracuda firewall.
As DNS the server use the ISP DNS.
On the positive side last night I did a restart of the server (didn't do it for over 4 months) and today the email is leaving the inbox in no time ...
Could that be that the server need a restart every now and then?
Those times look normal. About 14 seconds between receiving from client and relaying out. Keep an eye on logs during weeks ahead to see if this changes.
Feb 1 12:32:07 www postfix/qmgr[457]: 1D68515618EB: from=<xxx@xxx.com>, size=3030, nrcpt=1 (queue active) Feb 1 12:32:22 www postfix/smtp[13048]: 1D68515618EB: to=<xxx@xxxxx.ca>, relay=gmail.grafikom.ca[207.35.207.36], delay=29, status=sent (250 2.6.0 <C1E78E35.DD94%xxxx@xxxxx.com> Queued mail for delivery)
I removed the non-applicable lines.
13 seconds (but is missing original "connect from" line, so probably adds 16 secs?). Was this from a remote user (from=<xxxxx@xxxx.com>)?
Feb 1 12:13:43 www postfix/qmgr[457]: 9D0761561700: from=<xxxxx@xxxx.com>, size=5144, nrcpt=1 (queue active) Feb 1 12:14:25 www postfix/smtp[12810]: 9D0761561700: to=<xxxx@yahoo.com>, relay=a.mx.mail.yahoo.com[209.191.118.103], delay=63, status=sent (250 ok dirdel)
42 seconds (plus 20?). Similar comment to last one - remote user?
Feb 1 12:06:35 www postfix/qmgr[457]: D8DB41557739: from=<xxx@xxxx.com>, size=931, nrcpt=1 (queue active) Feb 1 12:07:00 www postfix/smtpd[12719]: connect from unknown[192.168.0.120] Feb 1 12:07:00 www postfix/smtpd[12719]: D7760156164F: client=unknown[192.168.0.120] Feb 1 12:07:00 www postfix/cleanup[12720]: D7760156164F: message-id=<978A9F1907FF3F46B7638745086B8CBA3AC43B@naexcmb4.na.global.ad> Feb 1 12:07:00 www postfix/qmgr[457]: D7760156164F: from=<xxxx@TMP.com>, size=11626, nrcpt=1 (queue active) Feb 1 12:07:00 www postfix/smtpd[12719]: disconnect from unknown[192.168.0.120] Feb 1 12:07:01 www postfix/pipe[12806]: D7760156164F: to=<xxxx@xxx.com>, relay=cyrus, delay=1, status=sent (xxxxxx.com) Feb 1 12:07:16 www postfix/smtp[12810]: connect to echoadvertising.com[64.74.223.9]: Operation timed out (port 25) Feb 1 12:07:16 www postfix/smtp[12810]: D8DB41557739: to=<xxxx@echoadvertising.com>, relay=none, delay=171983, status=deferred (connect to echoadvertising.com[64.74.223.9]: Operation timed out)
The message with ID D8DB41557739 is an outgoing message which is unable to be delivered and has been sitting in your queue for about 2 days.
The second message from your lan (D7760156164F: client=unknown[192.168.0.120]) took 1 second to go through.
Took 10 seconds in that attempt but has been sitting in outgoing queue for about 30 minutes - possibly due to greylisting on relaying server or some other reason why initial attempt couldn't get through. Maybe the relaying server is intentionally creating delays as an anti-spam measure - who knows.
If the guess about remote users is correct (only 1 obvious LAN message there), I would be looking to see if the pattern follows in other messages - does outgoing mail from LAN go through promptly? If problem seems to be remote users then I would be looking at their connection method. Also, do they connect direct to your server or does their traffic go through the Barracuda?
If your server is very busy and outgoing mail from LAN users is also generally delayed then maybe you need an upgrade to server (RAM?).
Apart from that, nothing else I can comment on since I'm only seeing a very small selection of log.
Feb 1 12:32:07 www postfix/qmgr[457]: 1D68515618EB: from=<xxx@xxx.com>, size=3030, nrcpt=1 (queue active) Feb 1 12:32:22 www postfix/smtp[13048]: 1D68515618EB: to=<xxx@xxxxx.ca>, relay=gmail.grafikom.ca[207.35.207.36], delay=29, status=sent (250 2.6.0 <C1E78E35.DD94%xxxx@xxxxx.com> Queued mail for delivery)
This is from my LAN going out
I removed the non-applicable lines.
13 seconds (but is missing original "connect from"
line, so probably adds 16 secs?). Was this from a
remote user (from=<xxxxx@xxxx.com>)?
Feb 1 12:13:43 www postfix/qmgr[457]: 9D0761561700: from=<xxxxx@xxxx.com>, size=5144, nrcpt=1 (queue active) Feb 1 12:14:25 www postfix/smtp[12810]: 9D0761561700: to=<xxxx@yahoo.com>, relay=a.mx.mail.yahoo.com[209.191.118.103], delay=63, status=sent (250 ok dirdel)
42 seconds (plus 20?). Similar comment to last one -
remote user?
Same here, my LAN going out.
Feb 1 12:06:35 www postfix/qmgr[457]: D8DB41557739: from=<xxx@xxxx.com>, size=931, nrcpt=1 (queue active) Feb 1 12:07:00 www postfix/smtpd[12719]: connect from unknown[192.168.0.120] Feb 1 12:07:00 www postfix/smtpd[12719]: D7760156164F: client=unknown[192.168.0.120] Feb 1 12:07:00 www postfix/cleanup[12720]: D7760156164F: message-id=<978A9F1907FF3F46B7638745086B8CBA3AC43B@nae xcmb4.na.global.ad> Feb 1 12:07:00 www postfix/qmgr[457]: D7760156164F: from=<xxxx@TMP.com>, size=11626, nrcpt=1 (queue active) Feb 1 12:07:00 www postfix/smtpd[12719]: disconnect from unknown[192.168.0.120] Feb 1 12:07:01 www postfix/pipe[12806]: D7760156164F: to=<xxxx@xxx.com>, relay=cyrus, delay=1, status=sent (xxxxxx.com) Feb 1 12:07:16 www postfix/smtp[12810]: connect to echoadvertising.com[64.74.223.9]: Operation timed out (port 25) Feb 1 12:07:16 www postfix/smtp[12810]: D8DB41557739: to=<xxxx@echoadvertising.com>, relay=none, delay=171983, status=deferred (connect to echoadvertising.com[64.74.223.9]: Operation timed out)
The message with ID D8DB41557739 is an outgoing
message which is unable to be delivered and has been
sitting in your queue for about 2 days.
The second message from your lan (D7760156164F:
client=unknown[192.168.0.120]) took 1 second to go
through.
Took 10 seconds in that attempt but has been sitting
in outgoing queue for about 30 minutes - possibly due
to greylisting on relaying server or some other
reason why initial attempt couldn't get through.
Maybe the relaying server is intentionally creating
delays as an anti-spam measure - who knows.
If the guess about remote users is correct (only 1
obvious LAN message there), I would be looking to see
if the pattern follows in other messages - does
outgoing mail from LAN go through promptly? If
problem seems to be remote users then I would be
looking at their connection method. Also, do they
connect direct to your server or does their traffic
go through the Barracuda?
If your server is very busy and outgoing mail from
LAN users is also generally delayed then maybe you
need an upgrade to server (RAM?).
Apart from that, nothing else I can comment on since
I'm only seeing a very small selection of log.
-david
the LAN users are going out directly, remote users are going thru the Barracuda firewall.
It can be RAM, I only have 1 GB of RAM in that machine and is also serving AFP out to about 20 client.
