eximX0902w at linuxwan
Feb 22, 2009, 2:16 AM
Post #6 of 6
(1261 views)
Permalink
|
|
[Bug 811] With SRS Exim crashes with SIGSEGV
[In reply to]
|
|
------- You are receiving this mail because: -------
You are on the CC list for the bug.
http://bugs.exim.org/show_bug.cgi?id=811
--- Comment #6 from Ted Cooper <eximX0902w [at] linuxwan> 2009-02-22 10:16:08 ---
I have been unable to replicate problem here so you will have to run the
commands there and then post the results.
Just a few things before we go over the edge - do you have all the following
options set in the configuration file? If setting them fixes the problem I
would really love to know so I can target that part of the code. Also, the SRS
router should look something like this.
If tweaking those settings fails to stop it seg faulting, then we'll need to
grab the core file and compiled binary.
# srs global config options
hide srs_secrets = somereallysecretkey
srs_maxage = 30
srs_hashlength = 6
srs_hashmin = 6
srs_usetimestamp = yes
srs_usehash = yes
# srs router
srs_fwd:
driver = redirect
condition = ${if !def:srs_db_key}
domains = +local_domains
data = $local_part
srs = forward
srs_alias = iutbeziers.fr
check_ancestor
qualify_preserve_domain
no_verify
First, we need to separate the delivery process so we can target that part. You
probably don't want to do this on a production server. If you can, run it on a
different port (24) and either use that as your SMTP AUTH host or setup a
manual route to deliver to it.
Run exim in queue only mode with no queue runners and deliver a mail to it.
/path/to/exim -bd -odq -C /path/to/special/configuration
You'll want either stop your normal exim daemon or change where this special
config saves its spool to (spool_directory = /path/to/special/spool.
Deliver the message initially and have it sitting in the queue. It will be
given a message id which you can look for with /path/to/exim -C
/path/to/special/configuration -bp
To trigger the delivery and possibly get a core file run
ulimit -c unlimited
strace /path/to/exim -M <msg id> > output.txt
If that seg faults it will give you a core.<pid> file (or just core). This file
will contain the memory contents of the exim process and may contain privileged
information so you will probably not want to post it on bugzilla .. it will
also be many Mb in size.
We'll try to avoid sending it for now, but this is going to be a pain :P
Post the output.txt on here, or possibly email it directly to me so you don't
have to go through redacting bits.
So you don't have to send the core file, exim and every other compiled library
linked to the exim binary to me, I'll need you to post the stack trace of the
segfault too:
gdb /path/to/exim /path/to/core/file
> bt
The bt command will print the back trace.
Post all that info and hopefully I'll have some idea of what is going on
--
Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email
--
## List details at http://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##
|
