Re: [gentoo-server] mailman^postfix update problems

[Christian Parpart]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Monday 21 June 2004 20:07, Joby Walker wrote:
> Probably a stale pid file.
>
> 1) Stop the daemon, disable any cron jobs and access to the website
> 2) Check <mailmanroot>/locks/, it should be empty.  If not...
> 3) delete all files, except the .keep.
> 4) Re-enable cron jobs and website access, restart the daemon.

Hi again,

now (after a full backup), I completely unmerged mailman, updated postfix to 
the current release, tested it (it's running fine), and now emerge mailman 
again.

CLEANUP
# cd /usr/local
# /etc/init.d/mailman stop
# /etc/init.d/apache stop
# tar xjpf mailman-backup.tbz2 mailman
# emerge -C mailman
# rm -rf mailman

INSTALL:
# emerge mailman
# cd mailman
# find . -name .keep -exec rm {} \;
# bin/check_perms -f
# echo >>Mailman/mm_cfg.py <<EOF
DEFAULT_EMAIL_HOST = 'lists.surakware.net'
DEFAULT_URL_HOST = 'lists.surakware.net'
DEFAULT_URL_PATTERN = 'http://%s/mailman/'
MTA = 'Postfix'
DELIVERY_MODULE = 'SMTPDirect'
POSTFIX_STYLE_VIRTUAL_DOMAINS = ['lists.surakware.net']
add_virtualhost(DEFAULT_URL_HOST, DEFAULT_EMAIL_HOST)
EOF
# su - mailman
# bin/newlist mailman
# bin/genaliases
# vim /etc/postfix/main.cf      # (ensure contens as below)
# su -c '/etc/init.d/postfix start'
# su -c '/etc/init.d/mailman start'
# ls -l logs
total 4
- -rw-rw-r--  1 mailman mailman   0 Jul  3 22:10 error
- -rw-rw-r--  1 mailman mailman 672 Jul  3 22:20 qrunner
# cat logs/qrunner
Jul 03 22:20:18 2004 (3813) BounceRunner qrunner started.
Jul 03 22:20:18 2004 (3812) ArchRunner qrunner started.
Jul 03 22:20:18 2004 (3814) CommandRunner qrunner started.
Jul 03 22:20:18 2004 (3815) IncomingRunner qrunner started.
Jul 03 22:20:18 2004 (3816) NewsRunner qrunner started.
Jul 03 22:20:19 2004 (3817) OutgoingRunner qrunner started.
Jul 03 22:20:19 2004 (3818) VirginRunner qrunner started.
Jul 03 22:20:19 2004 (3819) RetryRunner qrunner started.
Jul 03 22:20:20 2004 (3810) Master qrunner detected subprocess exit
(pid: 3817, sig: 11, sts: None, class: OutgoingRunner, slice: 1/1) 
[restarting]
Jul 03 22:20:20 2004 (3821) OutgoingRunner qrunner started.
# bin/newlist test cparpart@×××××××××.net test
# tail logs/qrunner
Jul 03 22:20:18 2004 (3816) NewsRunner qrunner started.
Jul 03 22:20:19 2004 (3817) OutgoingRunner qrunner started.
Jul 03 22:20:19 2004 (3818) VirginRunner qrunner started.
Jul 03 22:20:19 2004 (3819) RetryRunner qrunner started.
Jul 03 22:20:20 2004 (3810) Master qrunner detected subprocess exit
(pid: 3817, sig: 11, sts: None, class: OutgoingRunner, slice: 1/1) 
[restarting]
Jul 03 22:20:20 2004 (3821) OutgoingRunner qrunner started.
Jul 03 22:26:25 2004 (3810) Master qrunner detected subprocess exit
(pid: 3821, sig: 11, sts: None, class: OutgoingRunner, slice: 1/1) 
[restarting]
Jul 03 22:26:25 2004 (5113) OutgoingRunner qrunner started.
# 

====== part of postfix/main.cf begin
mydestination = $myhostname, $mydomain, localhost.$mydomain,
    mail.$mydomain, chat.$mydomain, sws.informatik.hu-berlin.de

alias_maps = hash:/etc/mail/aliases, hash:/usr/local/mailman/data/aliases

virtual_alias_domains = lists.surakware.net
virtual_alias_maps = hash:/usr/local/mailman/data/virtual-mailman

owner_request_special = no
recipient_delimiter = +
unknown_local_recipient_reject_code = 550
====== part of postfix/main.cf end

So.... I conclude, that the problem is not really in postfix, and not really 
because I really found some lost lock files in mailman/locks before.
However, anytime a new mail is to be sent out, a subprocess of qrunnier dies, 
and in my /var/log/messages I can find something like:

Jul  3 22:26:25 sws postfix/smtpd[5112]: connect from localhost[127.0.0.1]
Jul  3 22:26:25 sws postfix/smtpd[5112]: lost connection after CONNECT from 
localhost[127.0.0.1]
Jul  3 22:26:25 sws postfix/smtpd[5112]: disconnect from localhost[127.0.0.1]

How now can I solve this problem?

Many many thanks,
Christian Parpart.

> Christian Parpart wrote:
> > Joby Walker said:
> >>Christian,
> >>
> >>Have you tried manually running check_perms?  Have you checked the
> >>mailman logs?
> >
> > I've invoked check_perms several times. even check_db.
> >
> >>The mailman logs are pretty good, I usually find some
> >>important info on why mailman misbehaves for me.
> >
> > in file /usr/local/logs/qrunner:
> > [...]
> > Jun 21 17:48:10 2004 (3953) ArchRunner qrunner started.
> > Jun 21 17:48:10 2004 (3954) BounceRunner qrunner started.
> > Jun 21 17:48:10 2004 (3956) IncomingRunner qrunner started.
> > Jun 21 17:48:10 2004 (3957) NewsRunner qrunner started.
> > Jun 21 17:48:10 2004 (3955) CommandRunner qrunner started.
> > Jun 21 17:48:11 2004 (3959) VirginRunner qrunner started.
> > Jun 21 17:48:11 2004 (3960) RetryRunner qrunner started.
> > Jun 21 17:48:11 2004 (3958) OutgoingRunner qrunner started.
> > Jun 21 17:48:11 2004 (3951) Master qrunner detected subprocess exit
> > (pid: 3958, sig: 11, sts: None, class: OutgoingRunner, slice: 1/1)
> > [restarting]
> > [...lots.of.repititions...]
> > Jun 21 17:48:13 2004 (3972) OutgoingRunner qrunner started.
> > Jun 21 17:48:14 2004 (3951) Master qrunner detected subprocess exit
> > (pid: 3972, sig: 11, sts: None, class: OutgoingRunner, slice: 1/1)
> > [restarting]
> > Jun 21 17:48:14 2004 (3951) Qrunner OutgoingRunner reached maximum
> > restart limit of 10, not restarting.
> >
> > But which so called 'subprocess' exited?
> > And - more important - why?
> >
> > /usr/local/mailman/log/error seems untouched since Jun 10 20:05:01 2004.
> >
> > It must be anything regarding to the SMTP sending mechanism in /
> > or related to mailman. But sending mails via squirrelmail works
> > fine, so this (I guess) can't be really the reason.
> >
> > Thanks in advance,
> > Christian Parpart.
> >
> >>Christian Parpart wrote:
> >>>Hi all,
> >>>
> >>>I recently updated my postfix and mailman resulting into some mailman
> >>>issues,
> >>>that is, I can send mails to the mailinglists it serves, they get
> >>>archived,
> >>>but they won't be sent out to the subscribers. Why?
> >>>
> >>>I already look into postfix' main.cf and the Mailman/Default.py for some
> >>>changes, but nothing special found.
> >>>
> >>>Also, the mailman ebuild seems to have some problems, it creates the
> >>>.keep
> >>>files, and then executes bin/check_perms which interprets even those
> >>>files as
> >>>directories what - of course - results into an error while emerging.
> >>>
> >>>I inserted an
> >>>
> >>>find . -type d -name .keep -exec rm -f {} \;
> >>>
> >>>above the check_perms command line in this ebuild.
> >>>
> >>>Thanks in advance,
> >>>Christian Parpart.

- -- 
 21:51:30 up 19 days, 11:20,  4 users,  load average: 0.40, 0.39, 0.42
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA5xuvPpa2GmDVhK0RAgQUAKCGhu+5vxL9uODqPWfbmFXDUMmx8wCcC4X0
EVNhXUJfbCF7h0c+7LDGt1c=
=a23W
-----END PGP SIGNATURE-----