Hello Tim,
>> I have tried to set up TMDA (redHat Linux 7.1, Sendmail,
>> Python2.2.2) and came to a stage where I can't understand what's goind
>> on.
>>
>> .procmailrc:
>>
>> # Set the necessary environment variables.
>> EXTENSION="$1"
>> :0
>> * EXTENSION ?? .
>> {
>> DELIMITER="+"
>> }
>> RECIPIENT="[EMAIL PROTECTED]"
>> SENDER=`formail -x Return-Path | sed 's/[<>]//g;s/^[ ]*//'`
>> # Run the message through tmda-filter.
>> :0 fW
>> | /usr/local/tmda-0.84/bin/tmda-filter -p
>>
>> :0 e
>> {
>> # If tmda-filter returned 99, don't set EXITCODE!
>> :0
>> ? /bin/test $? -eq 99
>> /dev/null
>>
>> :0
>> EXITCODE=$?
>> /dev/null
>> }
>>
>> # TMDA takes care of final delivery
>> DEFAULT=/dev/null
TL> First of all, it doesn't look as if you need all this stuff in your
TL> .procmailrc, so why not simplify it? Don't use procmail's 'f' flag
TL> (just ":0 W"), don't use tmda-filter's "-p" flag and delete the entire
TL> ":0 e" recipe. Simply run tmda-filter and then set DEFAULT.
OK, I will simplify the .procmailrc, but the problem is
somewhere else.
>> I had to create the ~/.tmda/pending (I didn't find any hints
>> on .tmda contents in installation document),
TL> tmda-filter creates ~/.tmda/pending the first time it needs it.
In my case it didn't. Instead, in ~/TMDA_DELIVERY_FAILURE it
was written that TMDA can't find the 'pending' directory and aborts.
>> From maillog:
>>
>> Sep 11 00:54:36 zaurum sendmail[7317]: h8B7sYc07315: SYSERR(root): mailer prog died
>> with signal 11
>> Sep 11 00:54:36 zaurum sendmail[7317]: h8B7sYc07315: to="| /usr/bin/procmail",
>> ctladdr=<[EMAIL PROTECTED]> (502/503), delay=00:00:01, xdelay=00:00:00,
>> mailer=prog, pri=30019, dsn=4.0.0,
>> stat=Deferred: prog mailer (/usr/sbin/smrsh) exited with EX_TEMPFAIL
TL> When TMDA runs into an unrecoverable problem, it exits with errorcode
TL> 75 (EX_TEMPFAIL). It writes a Python stack trace into LOGFILE_DEBUG
TL> (if you have defined that) or ~/TMDA_DELIVERY_FAILURE. If you can
TL> post the contents of that file, we can see what's dying.
It doesn't create this file! All I have are the two records in
maillog. Nothing else. If the ~/TMDA_DELIVERY_FAILURE were available,
I would continue to investigate the problem myself.
TMDA simply dies, leaving nothing to judge why. This is why I
asked what I should do to force TMDA to create as much trace info as
possible.
>> The previous problem was 'error 99' (cured by modifying
>> .procmailrc the way it looks now).
TL> You won't get an exit code of 99 if you don't use the -p flag.
In the original setup, I didn't use '-p' flag. Sendmail has
complained that the mailer has returned exit code 99.
>> How can this current problem be solved (dying with signal 11)?
TL> It depends on what's causing the signal 11. The stack trace may
TL> provide us with more information.
The problem is to create the stack trace dump.
>> What should I place into ~/.tmda/config to receive more
>> sensible diagnostics?
TL> You can set LOGFILE_DEBUG to a file of your choosing. The user
TL> running TMDA must be able to write to that file. Usually we suggest
TL> something like ~/.tmda/debug.log or ~/.tmda/logs/debug. You can name
TL> it whatever you like. TMDA doesn't write into the system mail log.
OK, this is perhaps the piece of advice I can really use. Thank
you.
The major problem, for me, presonally, with TMDA is that the
installation directions are scattered over many places of
documentation. That is about 'pending' directory etc. A sample setup,
including the contents of all control files, wouldn't harm.
With all respect,
Konstantin
_____________________________________________
tmda-users mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-users
