[Mailman-Users] database error message when upgraded to 2.1.2; CPU PROBLEM SOLVED!!!
NOW Website Coordinator
webperson at now.org
Fri May 2 23:59:04 CEST 2003
I had only turned off the MTA (Postfix) on the old server, and it was
helping me prove that it was Mailman running out of control.
It turns out that the problem (solved by Barry Warsaw!!) was a setting in
Postfix that caused bad, local addresses to be retried over and over and
over and over...
From Postfix's main.cf:
# The default setting is 550 (reject mail) but it is safer to start
# with 450 (try again later) until you are certain that your
# local_recipient_maps settings are OK.
#unknown_local_recipient_reject_code = 550
unknown_local_recipient_reject_code = 450
Once I changed it to 550, restarted Postfix, and then (this part is
important) deleted some (bad local addresses) messages from the out qfiles
folder, restarted mailman.... the CPU usage went down to around 0% from 98%.
Thanks everyone for the feedback.
>But the corollary of a message being output by MHonarc when mailmanctl was
>started is: was this caused by Mailman handing off a message for archiving
>to MHonarc for archiving and did this provoke MHonarc's warning message?
>If so, where did the message come from if (as I think your are saying
>below) the incoming MTA is not running and delivering messages to Mailman?
>Also, is there any sign of messages being inserted into the MHonarc
>archives which appear to be associated with the MM-MHonarc interaction at
>mailmanctl start time? What happens if you re-enable the archiving scheme
>and restart mailmanctl? Does the warning message come back too?
>>I decided to use the Qrunner command and try and run one runner at a
>>time. Here's what I did, I just commented out one runner each and ran it:
>>QRUNNERS = [
>> ('ArchRunner', 1), # messages for the archiver
>># ('BounceRunner', 1), # for processing the qfile/bounces directory
>># ('CommandRunner', 1), # commands and bounces from the outside world
>># ('IncomingRunner', 1), # posts from the outside world
>># ('NewsRunner', 1), # outgoing messages to the nntpd
>># ('OutgoingRunner', 1), # outgoing messages to the smtpd
>># ('VirginRunner', 1), # internally crafted (virgin birth) messages
>>and the winner was .... the OutgoingRunner. The rest were very behaved,
>>but the Outgoing Runner took up over 70% of the CPU on a machine that is
>>sending *nothing* out. Postfix is shut down.
>That seems to demonstrate that the OutgoingRunner thinks it has work to do
>and load in not being generated by an Archiver.
>I assume Mailman is using SMTP to pass outgoing mail to your outgoing MTA?
>Are the outgoing and incoming MTAs one and the same?
>Are you saying that the outgoing MTA used by Mailman is not running and
>accepting mail on port 25? If so, this might offer an explanation if MM
>has outgoing mail to send.. Mailman can be fairly aggressive in trying to
>hand off mail to an SMTP server if it has outgoing mail to send. Mailman
>could just be looping hard trying to get a connection to the MTA in order
>to send mail out.
>As there any evidence of mail in the $prefix/qfiles/* directories? Is
>there no evidence of what is happening in any of MM's logs; the qrunner,
>error and smtp-failure logs for instance?
>>Tried it on the new Linux machine, and same result. Outgoing runner by
>>itself takes up 97% of my machine's CPU. Note I am doing a full
>>mailmanctl stop and start to test this.
>>Thanks for everyone's help. Any suggestions as to how to get the
>>outgoing runner to behave?
More information about the Mailman-Users