[Dovecot] Procmail died with Signal 11
Hi to all out there,
i know this is not the right mailinglist for this issue, but i?m hoping to get maybe here an answer to my problem. i have succesfully migrated 12.000 mailboxes on a debaian with washington imap to maildir on sles10 with dovecot. although i had some post-converting work to do, i can say that almost everything went fine. but there is one problem with my LDA. When postfix is about to deliver a lot of mail (user-db is ldap) procmail dies with signal 11. here is the error log:
Apr 3 08:19:17 mailvirtuell postfix/local[22367]: E42EF33A96:
to=root@fh-rosenheim.de, relay=local, delay=17, delays=0.08/0.11/0/17,
dsn=5.3.0, status=bounced (Command died with signal 11: "IFS=
&& exec
/usr/bin/procmail || exit 75 #Ja". Command output: procmail: Exceeded
LINEBUF )
I googled the whole inet for this problem and found a lot, but no solution seems to fit my problem. most articles point on this parameter for procmailrc "Linebuf". i found out that the default value is 2048k for message header reading. i doubled this value, with no effect. but i guess that a 4kb message header space must be enough to deal with. one hint last: starting value for ldap timeout was 10 secs. with this value the log went crazy like the above log-snip. i changed the timeout to 300 secs and thougt that was the trick, but there are still some process wich are diying after all.
i?m using postfix 2.5.1, procmail 3.22, dovecot 1.0.13, with sophos and amavis.
i would be really grateful for any ideas.
best regards Johannes Grimm
==============================================================
Hochschule für angewandte Wissenschaften - Fachhochschule Rosenheim
University of Applied Sciences
Rechenzentrum
Johannes Grimm
Hochschulstr.1
D-83024 Rosenheim
Tel: +49-8031/805-229
Fax: +49-8031/805-223
johannes.grimm@fh-rosenheim.de
==============================================================
At 11:23 AM +0200 4/10/08, Johannes Grimm wrote:
Hi to all out there,
i know this is not the right mailinglist for this issue,
That is correct. See http://www.procmail.org/era/lists.html to find one more suitable.
but i?m hoping to get maybe here an answer to my problem.
Your problem is a bug in procmail that is brought out by rules that are not adequately paranoid about their input.
--
Bill Cole
bill@scconsult.com
Thank you Bill for your quick response. I´ve already joined the procmail-list but it seems that no ones using it.... I´m already waiting day´s for a response....
You probably need to know that if a process is about to die, hundreds of the similar processes with the same mail-contents have already been succesfully delivered. but there must be a special circumstance that some processes still dies an the mail get lost. i have already tried procmail -t (which actually means that procmail dies softly if so, and the mail should get requed, but this functions is not working in my case)
==============================================================
Hochschule für angewandte Wissenschaften - Fachhochschule Rosenheim
University of Applied Sciences
Rechenzentrum
Johannes Grimm
Hochschulstr.1
D-83024 Rosenheim
Tel: +49-8031/805-229
Fax: +49-8031/805-223
johannes.grimm@fh-rosenheim.de
==============================================================
Bill Cole schrieb:
At 11:23 AM +0200 4/10/08, Johannes Grimm wrote:
Hi to all out there,
i know this is not the right mailinglist for this issue,
That is correct. See http://www.procmail.org/era/lists.html to find one more suitable.
but i?m hoping to get maybe here an answer to my problem.
Your problem is a bug in procmail that is brought out by rules that are not adequately paranoid about their input.
participants (2)
-
Bill Cole
-
Johannes Grimm