outgoing mail failing: unknown transport error

sneakyimp · 02-29-2016, 02:47 PM

I have an urgent problem. My server can no longer get mail delivered. I've looked in the postfix mail log (/var/log/mail.log) and see gobs upon gobs of errors like this one:

Code:

Feb 29 20:41:33 ip-10-239-138-208 postfix/pickup[27765]: 5836B629E: uid=33 from=<www-data>
Feb 29 20:41:33 ip-10-239-138-208 postfix/cleanup[26956]: 5836B629E: message-id=<cd096258d2b243209882aad90ea5223f@www.mydomain.com>
Feb 29 20:41:33 ip-10-239-138-208 postfix/qmgr[1170]: 5836B629E: from=<noreply@mydomain.com>, size=789, nrcpt=1 (queue active)
Feb 29 20:41:33 ip-10-239-138-208 postfix/error[27917]: 5836B629E: to=<breestuart@example.com>, relay=none, delay=0.02, delays=0.01/0/0/0.01, dsn=4.3.0, status=deferred (unknown mail transport error)

I googled around a bit for the 'unknown mail transport error' message and how I might resolve it and came across this page which suggests that one use this command to try and find out the issue:

Code:

sudo egrep '(warning|fatal|panic):' /var/log/mail.log | less

That command yields many lines, the first few of which are this

Code:

Feb 28 06:57:31 ip-10-239-138-208 postfix/smtpd[17655]: warning: dict_nis_init: NIS domain name not set - NIS lookups disabled
Feb 28 06:57:33 ip-10-239-138-208 postfix/qmgr[1170]: warning: private/smtp socket: malformed response
Feb 28 06:57:33 ip-10-239-138-208 postfix/qmgr[1170]: warning: transport smtp failure -- see a previous warning/fatal/panic logfile record for the problem description
Feb 28 06:57:33 ip-10-239-138-208 postfix/master[1163]: warning: process /usr/lib/postfix/smtp pid 17659 killed by signal 11
Feb 28 06:57:33 ip-10-239-138-208 postfix/master[1163]: warning: /usr/lib/postfix/smtp: bad command startup -- throttling
Feb 28 06:58:34 ip-10-239-138-208 postfix/master[1163]: warning: process /usr/lib/postfix/smtp pid 17701 killed by signal 11
Feb 28 06:58:34 ip-10-239-138-208 postfix/master[1163]: warning: /usr/lib/postfix/smtp: bad command startup -- throttling

I don't see any fatal or panic messages in this output and am wondering what might be the problem on my server. Any advice would be much appreciated.

Emerson · 02-29-2016, 03:28 PM

Quote:

A signal 11 error, commonly know as a segmentation fault, means that the program accessed a memory location that was not assigned to it. A signal 11 error may be due to a bug in one of the software programs that is installed, or faulty hardware.

It can be also corrupted executable or library causing this, corruption in turn may point to failing hardware.

sneakyimp · 02-29-2016, 05:06 PM

Quote:

Originally Posted by Emerson

It can be also corrupted executable or library causing this, corruption in turn may point to failing hardware.

Oh good grief! This machine is a VM hosted on AWS -- I'm not really sure how to test the underlying hardware or how to be sure that the VM is moved to some fresh hardware. Might a software update (apt-get update/upgrade) have caused it?

A reboot of the machine appears to have cleared up the problem for the time being. It's pretty disconcerting that I was unaware of the problem -- but I don't really see any way to notify myself if the server cannot send any email. Any thoughts on how to be notified of catastrophic mail failures would be appreciated.