Skip, Ran into the same problem myself. The difference you are reporting is the wrapper program in /etc/smrsh and in /home/mailman/mail/.
It's difficult to maintain separate copies or to update the copies as necessary. You should change the /etc/smrsh/wrapper entry into a Logical link in /home/mailman/mail/wrapper. That way when you update the /home/mailman/mail/wrapper program the same happens in the /etc/smrsh/wrapper. Imagine if you will having a separate mailman_wrapper file because the Majordomo Utility is still active on your server. Yet another level of complexity. This should put you back on the path. HTH. Skip Montanaro wrote: > Over the past week I have upgraded several bits and pieces of the machine > from which I run a small number of Mailman-administered mailing lists. Mail > still isn't flowing, the natives are getting restless, and I'm about to > throw my laptop out the window. I'm hopeful someone here can steer me in > the right direction. > > I am running Mandrake 8.1, sendmail 8.12.1 and mailman 2.0.8. Any time I > try to post a message to the mailing list I get the following barf from > smrsh: > > ----- The following addresses had permanent fatal errors ----- > "|/home/mailman/mail/wrapper mailowner cedu" > (reason: Service unavailable) > (expanded from: cedu-admin ) > > ----- Transcript of session follows ----- > smrsh: wrapper not available for sendmail programs > 554 5.0.0 Service unavailable > > /usr/sbin/smrsh was compiled with /etc/smrsh as its CMDDIR, verified with > the strings command: > > # strings -a /usr/sbin/smrsh | egrep /etc > /etc/smrsh > > /etc/smrsh contains a link to /home/mailman/mail/wrapper: > > # ls -l /etc/smrsh/wrapper > lrwxrwxrwx 1 root root 26 Mar 14 2001 /etc/smrsh/wrapper -> >/home/mailman/mail/wrapper* > > which, when dereferenced, seems fine to me: > > # ls -lL /etc/smrsh/wrapper > -rwxr-sr-x 1 skip mailman 39212 Dec 10 13:41 /etc/smrsh/wrapper* > > If I run the above wrapper command manually, it complains: > > # /home/mailman/mail/wrapper mailowner cedu > Failure to exec script. WANTED gid 12, GOT gid 0. (Reconfigure to take 0?) > > This I don't understand, because the program is clearly set gid to the > mailman group (which happens to be gid 70, not 12 or 0 - gid 12 is the mail > group). Mailman was configured as > > ./configure --with-mail-gid=mail --with-cgi-gid=nogroup > > I tried configuring with-mail-gid=mailman instead and got: > > ----- The following addresses had permanent fatal errors ----- > "|/home/mailman/mail/wrapper mailowner cedu" > (reason: 2) > (expanded from: cedu-admin ) > > ----- Transcript of session follows ----- > Failure to exec script. WANTED gid 70, GOT gid 12. (Reconfigure to take 12?) > 554 5.3.0 unknown mailer error 2 > > FWIW, /usr/sbin/sendmail is setgid smmsp (gid == 25). Should that be the > with-mail-gid? > > Before upgrading from Mandrake 7.2 to 8.1 I was happily using sendmail > 8.12.0 and mailman 2.0.6. After the upgrade I fell back to sendmail 8.11.6 > and my problems began. Today I installed sendmail 8.12.1 and upgraded to > mailman 2.0.8. > > Any suggestions appreciated... > > -- > Skip Montanaro ([EMAIL PROTECTED] - http://www.mojam.com/) > > ------------------------------------------------------ > Mailman-Users maillist - [EMAIL PROTECTED] > http://mail.python.org/mailman/listinfo/mailman-users -- Albert E. Whale - http://www.abs-comptech.com/aewhale.html ---------------------------------------------------------------------- ABS Computer Technology, Inc. - ESM, Computer & Networking Specialists Sr. Network, OpenView and Systems Consultant ------------------------------------------------------ Mailman-Users maillist - [EMAIL PROTECTED] http://mail.python.org/mailman/listinfo/mailman-users