RE: [Exim] exim 3.22 problem on solaris 2.6

Top Page
Delete this message
Reply to this message
Author: EXIM Users
Date:  
To: 'Gary Palmer', exim-users
Subject: RE: [Exim] exim 3.22 problem on solaris 2.6
I reported this to the list about 3 days ago (see my mail "stale/corrupt -D
files in spool/input" for details). We are seeing LOTS of these appearing,
at first we thought it might be a memory/disk corruption issue but after
moving the bulk of our mail over to another solaris 2.6 server we started
seeing exactly the same problem. Both are running exim 3.22 and are
generating these files about once every 3 minutes. There's nothing fancy in
our configuration, just straight mail routing between servers.

Looking out files they seem to contain parts of /etc/services, parts of the
message headers, and even parts of the SMTP conversation?! They aren't
entirely plain text either. Could this been some sort of memory dump?

Anyone got any suggestions as to how we can log this further?, as it's
difficult to determine which messages will cause this problem it's hard to
start truss'ing random exim processes...

Looks like we'll have to go back to 3.16 :(

Jon Paxton

-----Original Message-----
From: Gary Palmer [mailto:gjp@in-addr.com]
Sent: 05 June 2001 05:51
To: exim-users@???
Subject: [Exim] exim 3.22 problem on solaris 2.6



Hi,

We recently tried upgrading to Exim 3.22 (compiled with 32 bit version
of gcc 2.95.2) on our inbound mx servers and while it appeared to
work, we noticed that a number of odd files were gathering in the
spool directory. They're all -D files, and they're all roughly the
same size. e.g.

-rw-------   1 exim     exim      811008 May 29 07:23 ./B/154hbB-0004oH-00-D
-rw-------   1 exim     exim      778240 May 29 08:03 ./B/154iEB-0006eL-00-D
-rw-------   1 exim     exim      778240 May 29 08:28 ./B/154icB-00055e-00-D
-rw-------   1 exim     exim      778240 May 29 09:33 ./B/154jdB-0005O2-00-D
-rw-------   1 exim     exim      778240 May 29 10:06 ./B/154k9B-0001yw-00-D
-rw-------   1 exim     exim      778240 May 29 11:12 ./B/154lAB-0004iQ-00-D
-rw-------   1 exim     exim      778240 May 29 11:45 ./B/154lgB-0002PZ-00-D
-rw-------   1 exim     exim      778240 May 29 12:50 ./B/154mhB-0003TT-00-D
-rw-------   1 exim     exim      778240 May 29 12:58 ./B/154mpB-0006TZ-00-D
-rw-------   1 exim     exim      778240 May 29 13:04 ./B/154mvB-0000rJ-00-D
-rw-------   1 exim     exim      778240 May 29 13:23 ./B/154nDB-0007Cw-00-D
-rw-------   1 exim     exim      778240 May 29 13:56 ./B/154njB-0002jy-00-D


The odd thing comes when you look in one of the files. At the start,
there is part of a header. The message-id I find mentioned for the
message on the machine where the file is found is not in the logs
anywhere (main or reject). No panic logs were created. Then a small
snipped of the configuration file, then what appears to be a dump of
malloc'd memory used by exim itself.

Anyone seen anything like this before? It doesn't happen often
enough, and these machines are under a load which makes trussing the
exim listener & its children an extreme way of debugging the problem,
and I'm not sure of any other way of determining what is going on.
I've since backed back down to 3.16 (which didn't have this problem),
but I'd like to upgrade.

Thanks,

Gary


--
## List details at http://www.exim.org/mailman/listinfo/exim-users Exim
details at http://www.exim.org/ ##


----------------------------------------------------------------------
If you have received this e-mail in error or wish to read our e-mail
disclaimer statement and monitoring policy, please refer to
http://www.drkw.com/disc/email/ or contact the sender.
----------------------------------------------------------------------