[Mailman-Users] Life lock
jonc at haht.com
Thu Oct 11 23:48:12 CEST 2001
split the list into sub lists, use an umbrella list to send to them all.
Get a machine with a faster processor and better disk subsystem. The
problem could be due to some built in timeouts, and anything that speeds up
the process is going to aid you.
For now, you will have to use your sendmail logs to grep for the message id
(you should be able to play around and figure that out from the logs), and
then grep for "sent" out of the attempts in your log for that message.
Assume those folks got the email and create another list without those
Good Luck - Jon Carnes
----- Original Message -----
From: "Daniel Tschan" <tschan at puzzle.ch>
To: <mailman-users at python.org>
Sent: Thursday, October 11, 2001 2:46 PM
Subject: [Mailman-Users] Life lock
> Today i tried to send a mail to a mailman mailing list containg about
> 100'000 members. It failed miserably. After about 100 Mails the first
> bounces returned. Because the list was still locked by qrunner, mailowner
> blocked. Apparently it waited for the release of the lock using busy
> waiting. After about 20 bounces the load was so high that sendmail started
> rejecting connections, which caused qrunner to block. Although there is a
> time limit for qrunner it never quit. A perfect life lock. I stopped the
> whole thing, reduced the time limit for qrunner, increased the time
> qrunner starts and disabled sendmail load limitations. But this dind't
> really help, it only delayed the problem. And to make things worse some
> s received the mail multiple times.
> Now I sit here with a half sent mailing and some annoyed users. What can I
> do to send the rest of the mailing?
> Some facts about the environment:
> Red Hat Linux 7.1 on a dual Pentium III 733
> Kernel 2.4.4
> ReiserFS 3.6.25
> Python 2.2
> Sendmail 8.11.6
> Mailman 2.0.6
> Mailman-Users maillist - Mailman-Users at python.org
More information about the Mailman-Users