[exim-dev] [Bug 1810] New: Mangled queue following system-fi…

Top Page
Delete this message
Reply to this message
Author: admin
Date:  
To: exim-dev
Subject: [exim-dev] [Bug 1810] New: Mangled queue following system-filter unseen deliver retry
https://bugs.exim.org/show_bug.cgi?id=1810

            Bug ID: 1810
           Summary: Mangled queue following system-filter unseen deliver
                    retry
           Product: Exim
           Version: 4.86
          Hardware: All
                OS: Linux
            Status: NEW
          Severity: bug
          Priority: medium
         Component: Delivery in general
          Assignee: nigel@???
          Reporter: stevep@???
                CC: exim-dev@???


Following an upgrade to 4.86.2 due to CVE-2016-1531 on a couple of systems,
compiled from source, I ran into this bug using a system-filter with an 'unseen
deliver':

if first_delivery and not error_message 
  then
        unseen deliver root@???
endif


2016-03-23 08:25:48 1agDAv-0004rt-3k == root@??? <system-filter>
R=eximtrap T=smtp_eximtrap defer (-1): smtp transport process returned non-zero
status 0x000b: terminated by signal 11
2016-03-23 08:25:50 1agDKY-0005MT-18 == root@??? <system-filter>
R=eximtrap T=smtp_eximtrap defer (-1): smtp transport process returned non-zero
status 0x000b: terminated by signal 11

This was triggered when the remote system stopped accepted delivery due to
load, and would **freeze** the undelivered messages in the queue.

I patched 4.86.2 with:

http://git.exim.org/exim.git/commitdiff/dadff1d47e54962b0fdf98e8ce5cef42b6cb7fb5

which fixed the signal 11 termination, BUT appears to mangle the message in the
spool queue. When delivery is retried (as it now should be) the -D file has
been deleted from the spool:

2016-03-23 09:21:11 1aieyV-0002ei-IR <= scripts@???
H=cache1.internal.XXXX.com (pm.XXXX.com) [10.0.2.60] P=esmtps
X=TLS1.0:RSA_AES_256_CBC_SHA1:256 CV=no S=1788
id=message.5uDrkhiw4UqUwa9FnRiPbvmpQMk@???
2016-03-23 09:21:11 1aieyV-0002ei-IR == root@??? <system-filter>
R=eximtrap T=smtp_eximtrap defer (-53): retry time not reached for any host
2016-03-23 09:21:13 1aieyV-0002ei-IR => xxxxxx@??? R=dnslookup
T=remote_smtp H=mx4.hotmail.com [65.55.33.119]
X=TLS1.2:ECDHE_RSA_AES_256_CBC_SHA384:256 CV=yes DN="CN=*.hotmail.com" C="250
<message.5uDrkhiw4UqUwa9FnRiPbvmpQMk@???> Queued mail for delivery"
2016-03-23 09:22:10 1aieyV-0002ei-IR Completed
2016-03-23 09:22:22 1aiezJ-0002nA-Ow Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:22 1aiezJ-0002nJ-Qq Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:22 1aiezK-0002nq-5X Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:22 1aiezR-0002qy-QY Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:22 1aiezK-0002nb-1u Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:22 1aiezJ-0002nU-Vk Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:22 1aiezK-0002nm-48 Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:22 1aiezK-0002nX-6a Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:22 1aiezK-0002nw-7o Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:23 1aiezJ-0002nV-VR Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:23 1aiezJ-0002nW-Vo Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:23 1aiezK-0002oB-Eo Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:22:37 1aiezt-0002x4-EW Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:23:02 1aif0I-0002yj-9r Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:23:28 1aif0i-000302-Io Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:23:35 1aif0p-000307-Im Spool file 1aieyV-0002ei-IR-D not found
2016-03-23 09:24:22 1aif1G-00033i-8F Spool file 1aieyV-0002ei-IR-D not found

Although the above 1aieyV-0002ei-IR message id is marked as completed @09:22:10
on the first subsequent retry, the mail never arrives on eximtrap.XXXX.com and
the resulting 'not found' message reoccurs on queue reruns.

We were previously running a standard Debian pkg (no changes in config except
the addition 'keep_environment =' added), without issue.

I did attempt to reopen Bug 1671 - but I guess this is a different issue....

--
You are receiving this mail because:
You are on the CC list for the bug.