[Mailman-Users] Re: Mailman not responding....

[Jeff Shultz]

On Thu, 14 Sep 2000 00:37:34 +0400, Michael Tokarev wrote:

>Jeff Shultz wrote:
>> 
>> >  What other logs says?
>> 
>> All that syslog and mail log say is that postfix receives the message
>> and turns it around to "|/home/mail/wrapper mailcmd test"
>> Mailman's logs are, other than some stuff yesterday when I set up the
>> listserv, empty.
>
>So, postfix thinks that it successefully delivered message to mailman's
>wrapper, right?  Or it reports some other errors?

As long as the proper gid (99 aka nobody) is getting passed between the
two, postfix thinks it was successful. No other errors that I've found.


>BTW, what with directories in /home/mail -- maybe there are some
>symlink(s) here to /misc exists?

You mean /home/mailman /mail?  wrapper is the only file in that
directory. 
>

>Again, I asked for configuration files for the two products, not about
>general machine's config.
>
Um... that's on a separate machine from this one. Anything specific you
are looking for? 

>> >Why, at the end, your mail delivery wanted to do something with /misc
>> >automounted directory?!
>> 
>> I wish I knew... but I blew that stuff away and it doesn't appear to
>> have hurt anything.
>
>But it chagned somethings, as you stated already.
>
It stopped registering an error message. That's the only apparent
change. 

>Ok.  If mailman's wrapper returns 0 (success) when called by postfix
>(see above), then it is definitely a mailman's problem, or your mailman's
>setting's one.  I assume that this is the case -- postfix logs
>successeful delivery to mailman's wrapper.
>
>Please do the following.
>
>as root (or, better, as other user that is a member of group that mailman
>expects, -- i.e. that group with that wrapper called from postfix) execute:
>
>  # newgrp GROUP-MAILMAN'S-WRAPPER-WANTS
>     (substitute correct value here)
>  # /home/mail/wrapper mailcmd test < THE-MESSAGE-THAT-YOU-SENT-TO-IT-BEFORE
>     (watch for possible output here)
>  # echo $?
>     (should be 0)
>
program appeared to lock up or loop. Broke it with ctrl-c.

>Or, if nothing interesting will appear, add "strace -o trace" (without quotes)
>before /home/... (this is possible only for root, as wrapper is setgid) and
>watch the newly-created "trace" file carefully.

At last count it was at 146k... and I'm seeing a whole lot of (No such
file or directory) entries at the ends of lines. 

I am starting to get errors in the Mailman error log - but I have no
idea how to read these Tracebacks. I don't know Python.
Jeff Shultz
Shultz Infosystems
www.shultzinfosystems.com