time limit exceeded - Local mail delivery fails

Question

Level 1

95 points

time limit exceeded - Local mail delivery fails

I've recently had a problem (once last week and once again today) where postfix/cyrus seems to run fine (10.4 server latest software update) for a while and then with little pattern I can determine it will stop delivering email to local users.

The troubling thing is that when the problem starts I see these two lines consistently until the mail server is restarted:

postfix/pipe[24728]: warning: pipe commandwrite: write time limit exceeded
postfix/pipe[24728]: 27B6E3BFA60: to=<user@localmachine>, relay=cyrus, delay=1000, status=bounced (Command time limit exceeded: "/usr/bin/cyrus/bin/deliver")

Once the mail process is restarted (via Server Admin) the problem goes away and waits to sneak up another day.

A quick search on the web has been unproductive. Anybody seen this?

Logfile:

Oct 6 09:26:14 mailserver postfix/smtpd[24723]: connect from PUBLIC MAILSERVER
Oct 6 09:26:14 mailserver postfix/smtpd[24723]: 27B6E3BFA60: client=PUBLIC MAILSERVER
Oct 6 09:26:14 mailserver postfix/cleanup[24726]: 27B6E3BFA60: message-id=<1128615965.4345501dd3b13@webmail.ANOTHER-CORP-EMAIL-SERVER>
Oct 6 09:26:14 mailserver postfix/smtpd[24723]: disconnect from PUBLIC MAILSERVER
Oct 6 09:26:14 mailserver postfix/qmgr[89]: 27B6E3BFA60: from=<USER@HOST>, size=6002, nrcpt=1 (queue active)

[... routine mail stuff... - but no mail deliveries to local users....]

Oct 6 09:40:20 mailserver postfix/cleanup[1449]: 9882F3BFAE9: message-id=<OFA26D55BD.4BD0AC2B-ON88257092.005B94B0-88257092.005B94EA@xxxxxxxx>
Oct 6 09:40:20 mailserver postfix/qmgr[89]: 9882F3BFAE9: from=<unlvinformation@ccmail.nevada.edu>, size=7048, nrcpt=2 (queue active)
Oct 6 09:40:20 mailserver postfix/smtpd[1445]: disconnect from PUBLIC MAILSERVER
Oct 6 09:42:54 mailserver postfix/pipe[24728]: warning: pipecommandwrite: write time limit exceeded
Oct 6 09:42:54 mailserver postfix/pipe[24728]: 27B6E3BFA60: to=<user@localmachine>, relay=cyrus, delay=1000, status=bounced (Command time limit exceeded: "/usr/bin/cyrus/bin/deliver")
Oct 6 09:42:54 mailserver postfix/cleanup[9540]: 3B70A3BFAF8: message-id=<20051006164254.3B70A3BFAF8@localmachine>
Oct 6 09:42:54 mailserver postfix/qmgr[89]: 3B70A3BFAF8: from=, size=7846, nrcpt=1 (queue active)
Oct 6 09:42:54 mailserver postfix/qmgr[89]: 27B6E3BFA60: removed

Posted on Oct 6, 2005 6:27 PM

Reply

Answer 1

Oct 16, 2005 2:58 PM in response to joshb

I see the exact same thing. This has happened almost once every week, but not on the exact time every week so I do not think it is the log roll problem that existed on earlier versions.
I run 10.3.9 and the problem has followed me from 10.3.7.

Now I have followed the instructions in this article:
http://www.afp548.com/article.php?story=20050517092837520 to reconstruct the cyrus db to see if it helps.

Make sure to read the discussions below the article since the files have changed a bit since it was written.

I just made the changes today so I do not know if it fixes the problem.

Reply