NOTE added 2013-11-10: I think I found the cause. Read about it in this FreeBSD Forums post.
I’ve seen seeing the following in the mail log for my imap server.
Sep 27 00:00:00 imaps newsyslog[43463]: logfile turned over Sep 27 03:01:23 imaps postfix/pickup[48680]: 3E479796978: uid=0 from=<root> Sep 27 03:01:23 imaps postfix/cleanup[53346]: 3E479796978: message-id=<20130927030123.3E479796978@imaps.example.org> Sep 27 03:01:23 imaps postfix/qmgr[71070]: 3E479796978: from=<root@imaps.example.org>, size=1157, nrcpt=1 (queue active) Sep 27 03:01:23 imaps postfix/cleanup[53346]: 45C5C796979: message-id=<20130927030123.3E479796978@imaps.example.org> Sep 27 03:01:23 imaps postfix/local[53348]: 3E479796978: to=<root@imaps.example.org>, orig_to=<root>, relay=local, delay=0.05, delays=0.03/0.01/0/0.01, dsn=2.0.0, status=sent (forwarded as 45C5C796979) Sep 27 03:01:23 imaps postfix/qmgr[71070]: 45C5C796979: from=<root@imaps.example.org>, size=1299, nrcpt=1 (queue active) Sep 27 03:01:23 imaps postfix/qmgr[71070]: 3E479796978: removed Sep 27 03:01:23 imaps postfix/postdrop[53313]: warning: uid=0: File too large Sep 27 03:01:23 imaps postfix/smtp[53349]: 45C5C796979: to=<dan@example.org>, orig_to=<root>, relay=tallboy.example.org[10.0.0.1]:25, delay=0.24, delays=0/0.01/0.2/0.03, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as 8014C1C51B2D) Sep 27 03:01:23 imaps postfix/sendmail[53309]: fatal: root(0): message file too big Sep 27 03:01:23 imaps postfix/qmgr[71070]: 45C5C796979: removed
The mail messages in question are the daily run and security run emails that go out from FreeBSD machines. They aren’t big. As shown above, you can see one was only 1157 bytes.
What is curious about this setup: the imap server runs in a jail. The relay mail server is on the jail host.
From what I can tell, this is definitely not an email which is too large. Here are the configuration items from the sending and from the receiving postfix servers.
# postconf | grep -i size berkeley_db_create_buffer_size = 16777216 berkeley_db_read_buffer_size = 131072 body_checks_size_limit = 51200 bounce_size_limit = 50000 header_size_limit = 102400 mailbox_size_limit = 51200000 message_size_limit = 10240000 tcp_windowsize = 0
$ postconf | grep -i size berkeley_db_create_buffer_size = 16777216 berkeley_db_read_buffer_size = 131072 body_checks_size_limit = 51200 bounce_size_limit = 50000 header_size_limit = 102400 mailbox_size_limit = 51200000 message_size_limit = 10240000 tcp_windowsize = 0
What do you think is the problem?
In a perhaps futile attempt to fix this, I upgraded the sending postfix to 2.10.1, the same version as the receiver.
Q: How much space is left on the disk?
A: about 400G
sending system:
receiving system:
Q: Does sending email via telnet work?
A: Yes
And the logs:
The postfix upgrade was futile. The problem persists.
Yesterday, the weekly message got through. Today, the daily run message arrived. The security message did not.