I eventually traced that my mailman installation is delaying the
distribution of messages because it is trapped in a bunch of "bounce
processing", dealing with undelivered messages, etc... in this regard,
could I simply inactivate bounce processing? If I do so I should expect
an increase in speed and performance of mailman and a decrease in
delays, but I might risk that some members, whose messages could not be
delivered for whatever reasons, might loose some of these messages. Is
that so? Would you recommend setting Mailman to not performing bounce
processing to improve its performance?
I'm using Mailman 2.1.14 on my congregation's website, hosted by
site5.com. We use it for a couple of dozen mailing lists, and haven't
had many troubles.
Problem 1: someone noticed that all the list archives were publicly
viewable, and even Googleable, which is a Bad Thing. So, I went into
the Mailman admin interface to turn off the archives.
In the Archiving Options page, I found the "archive_private" setting.
The description was a bit vague, but it seemed to be what I wanted,
and currently was set to "public". So, I set it to "private", and
Problem 2: boom! The Admin interface crashed, saying
>Bug in Mailman version 2.1.14
>We're sorry, we hit a bug!
>Please inform the webmaster for this site of this problem. Printing
>of traceback and other system information has been explicitly
>inhibited, but the webmaster can find this information in the
>Mailman error logs.
Problem 3: the Mailman admin interface for this list seems to be
permanently borked. When I go back to the admin page, I'm invited to
log in, but when I do it crashes again with the above message.
Restarting Firefox (8.0.1 on OS X 10.5.8), deleting cookies, using a
completely different browser: nothing helps.
Small good news 1: the list is still processing email. Small good
news 2: the other lists' admin interfaces on this machine aren't
borked, just the two I tried setting archive_private to "private" on.
(I broke the second one so I could test whether emails were being
So, I need help on two items:
1) How do I un-bork those two mailing lists, without losing all of
the list settings (especially the list of members)?
One note: since this is a shared server, my account is nailed down
pretty tightly, and I'm pretty sure I don't even have access to the
main mailman directories. So, I'll probably need to pass instructions
on to the hosting company's tech support.
2) How can I keep the public from seeing these archives? In truth we
don't need them even privately, so either making them private or
getting rid of them completely would be fine by me.
Of course, if the answer is "set archive_private to private", then I
need help on a third item:
3) How can I debug Mailmain's problem with the archive_private setting?
Daniel T. Griscom griscom(a)suitable.com
Suitable Systems http://www.suitable.com/
1 Centre Street, Suite 204 (781) 665-0053
Wakefield, MA 01880-2400
Thanks Ralf, I understand.
Apologies If I did not make myself understandable. I am not sending garbage to non-existing addresses. All addresses are verified and existing but not all messages appear to be delivered equally well, to my surprise, while some servers appear to refuse messages coming from this list for a while, until they are accepted for distribution, thus delaying overall mailman performance while it keeps trying to deliver, therefore I was considering to cancel bounce processing to speed up mailman performance. When I ask my hosting provider for the SMTP logs there appears to be activity by mailman, bounce processing, when there are no messages to distribute through the list, therefore I inferred that we are experiencing delays in distribution, beyond 30-60 minutes, from a message being sent to a message being distributed, because of this bounce processing actions. This is a list with about 600 members.
According to your answer this would not work this way.
* Lluis Montoliu<montoliu(a)cnb.csic.es>:
> Dear listers,
> I eventually traced that my mailman installation is delaying the
> distribution of messages because it is trapped in a bunch of "bounce
> processing", dealing with undelivered messages, etc... in this
> regard, could I simply inactivate bounce processing?
No. You need to clean out your lists. Why are you sending out garbage to
non-existing addresses at? A properly maintained list has a low
percentage of undeliverables - when (as in your case) the undeliverables
are exceeding the normal fallout of about 3%, then you're having bad
> If I do so I should expect an increase in speed and performance of
> mailman and a decrease in delays, but I might risk that some members,
> whose messages could not be delivered for whatever reasons, might loose
> some of these messages. Is that so?
> Would you recommend setting Mailman to not performing bounce processing
> to improve its performance? thanks Lluis
Sending mail to non-deliverable adresses is going to get you into
blacklists. That's how Hotmail and SenderScore work. Furthermore
you're going to hit spamtraps sooner or later.
-- Ralf Hildebrandt Geschäftsbereich IT | Abteilung Netzwerk Charité -
Universitätsmedizin Berlin Campus Benjamin Franklin Hindenburgdamm 30 |
D-12203 Berlin Tel. +49 30 450 570 155 | Fax: +49 30 450 570 962
ralf.hildebrandt(a)charite.de | http://www.charite.de
Dr. Lluis Montoliu
Investigador Cientifico - Research Scientist CSIC
Centro Nacional de Biotecnologia (CNB-CSIC)
Campus de Cantoblanco
C/ Darwin, 3
28049 Madrid (Spain)
Tel. +34-91-5854844 / Fax +34-91-5854506
U756 CIBERER: http://www.ciberer.es
Spanish EMMA node: http://www.emmanet.org
A list owner of one of our 500+ mailing lists notified me that
postings from their non-archived list was being harvested by
archiveorange.com and could be found via Google. I found that many of
our lists are archived on this site, and this is not something that we
really want. I am just querying the list to see if this has happened
with others using Mailman. The website explains that archiveorange
"subscribes" to lists and then harvests messages. It seems that the
only way this could happen to a non-archived list is for the mailbox
directories to readable to the outside. Is there something else that
is going on here?
Thanks for any information that anyone has to offer.
I searched and looked, but am unable to find where the text for the
message that is sent to a rejected moderation from a non-member
("Non-members are not allowed to post messages to this list") is stored
so I can change it to something a tad less cryptic.
Members of our list are also members of other, somewhat
topically-related, lists and sometimes cc those other lists. When a
member of the "other" list replies, it is held by our list since that
individual is NOT a member of our list. I can easily recognize when this
occurs, but the default reject message is a bit harsh for our purposes,
which are to be somewhat inclusive rather than walled off...
Something like "You must be a member of the xyz_group to post messages
to the list" (which is what I have manually inserted in place of the
default message), if I could but find when the terse message is stored.
Jack Hill, W4KH - BoatAnchors Listowner/Archiver
"Plus ca change, plus c'est la meme chose"
"Il n'y a que les idiots qui ne changent jamais d'idee"
Last week, we ran into a rather odd event (first time we've seen it in a
decade or so of Mailman ops), where a bounced message kept writing to its
/var/mailman/data/bounce-events-PID.pck file, until it filled the disk --
it wasn't a series of bounces (same delivery times on every received
header), it was the same bounce written over and over again (about 5000
times), unfortunately, that bounce contained a 7 MB base64 encoded
I thought I'd toss the details out here, to see if anyone had a thought on
what caused this, or (other than implementing a disk watching cron job,
which we already have), how to avoid it in the future ... though it
certainly seems like a fluke.
Anyhow, the message was first sent to the list at 16:41:35 (after manual
list owner approval of the moderated posting), and was posted normally at:
/var/mailman/logs/post:Nov 16 16:41:35 2011 (1560) post to [somelist]-news
from [someone](a)my.org size=7705563,
The message had a PDF attachment, and was from the list owner. The list had
about 120 members, and most of the members (there were some other
undeliverable addresses) got the message. The list owner got the message.
There appeared to be nothing bogus about the attachment.
The first bounce seemed normal:
Delivery failed for the following reason:
Server entsrmbx01.teledyne.com[184.108.40.206] failed with: 550 #5.1.0
Address rejected <someone>@teledyne.com
This has been a permanent failure. No further delivery attempts will be
... and then the first bounce was written to the
/var/mailman/data/bounce-events pck file at:
Received from localhost (localhost [127.0.0.1])
by [myserver] (Postfix) with ESMTP id 3272D601E21
for <[somelist]-news-bounces@myserver>; Wed, 16 Nov 2011 16:41:55
... but was written another 5000 times with it's PDF attachment.
... at 17:43 local, the disk filled with:
Nov 16 17:43:13 [hostname] postfix/smtpd: NOQUEUE: reject: MAIL from
localhost[127.0.0.1]: 452 4.3.1 Insufficient system storage; proto=ESMTP
Nov 16 17:43:13 [hostname] postfix/smtpd: warning: not enough free
space in mail queue: 36216832 bytes < 1.5*message size limit
Oddly, the bounce-events file's last write was considerably earlier:
-rw-rw-r-- 1 mailman mailman 37867700224 2011-11-16 16:52
The only questionable log entries I can find after/around the message
posting is for an abort of the Archrunner and Bouncerunner before the disk
Nov 16 17:06:08 2011 (1533) Master qrunner detected subprocess exit
(pid: 1555, sig: 9, sts: None, class: ArchRunner, slice: 1/1) [restarting]
Nov 16 17:06:14 2011 (27568) ArchRunner qrunner started.
Nov 16 17:06:18 2011 (1533) Master qrunner detected subprocess exit
(pid: 1556, sig: 9, sts: None, class: BounceRunner, slice: 1/1) [restarting]
Nov 16 17:06:20 2011 (27573) BounceRunner qrunner started.
Since the [somelist]-news list in question doesn't have archiving enabled
(never has), so I the ArchRunner restart. There were no other qrunner
restarts since 6:33 that morning, and the only later ones were when admin
staff went in to clear things up.
- 16:41 first bounce occurs and is written to bounce-events by PID 1556
- 16:52 PID 1556 makes (according to the file system) its last write to
- 17:02 ArchRunner and BounceRunner get restart signals, PID 27573 is the
- 17:43 the disk fills to the point postfix starts complaining
... that doesn't add up unless one presumes PID 1556 hung around and kept
writing bounce-events-1556.pck from 17:02 to 17:43 (and probably beyond).
Does the above add up to anything for anyone else? Is there something else
that should be checked?
Thanks for any insight.
Richard Haas <rhaas(a)rhaas.us>
GnuPG public key ID: 1CB7F0E2
Please forgive me if my question is a display of ignorance. I've only
very recently switched from ezmlm-idx to mailman and am still trying to
get my head around the difference in concept. I've searched the FAQs
and archives but wasn't actually too sure what exactly to use as search
I have noticed that Mailman identifies lists by there names only but
allows them to run in different mail domains. I realise that I can also
have different domains and host names for the web interfaces. However,
what I cannot find out for sure is if a site admin only exists per
Mailman install or also per mail domain. In other words, can I have an
admin who is allowed to create new lists (something I understand to be
exclusively a site admin task) for a certain mail domain but not on
other mail domains.
If I was running lists
and they were reachable on test(a)lists.example.com,
test1(a)lists.example.net, test2(a)lists.example.org could I have one
account which can create lists in example.com, another one which can
create lists in example.net and still another one which can create
lists in example.org?
Also, is there a trick to have lists test3(a)lists.example.net and
test3(a)lists.example.com simultaneously without needing to work with
I have installer last maiman on debian lenny, by apt-get.
There are 404 errors for interfaces on a domain.
There are mail warnings :
Nov 18 13:38:02 <server> postfix/postmap: warning:
/var/lib/mailman/data/virtual-mailman.db: duplicate entry: "<list>@<domain>"
(and other aliases)
/var/lib/mailman/data/aliases seems ok.
But /var/lib/mailman/data/virtual-mailman has a problem :
there are several lists of <domain> which are not deleted.
I tried to delete and recreate the list, same problem. And same problem
with all lists on that domain.
After reinstalling mailman, I have almost the same problem.
Now, there are no errors with virtual-mailman.db. But if I delete a list
of this domain, it stays in virtual-mailman. For others domains, no problem.
I have found that web admin/users interfaces are https URLs for the
domain mail of this list, and not http URL.
And the links from admin/users interfaces are https URLs, and coded as
In mm_cfg.py, there is
DEFAULT_URL_PATTERN = 'http://%s/cgi-bin/mailman/'
And no https directive in <domain>.vhost
Where can I search ?
We run Mailman as a department. Our core mail service is Exchange 2010.
Exchange is running in Shadow Redundancy mode which means it uses a Message DE duplication system.
I'm lead to believe that this is based on the Message-ID. (Please correct me if I'm wrong !)
In short, if a user posts to more than 1 list, Exchange dedupes the message on the way back.
This means users only get 1 list message. It's causing Chaos !
RFC's and the non PC tampering of message headers aside:
How can I use Mailman (or even sendmail) to remove the Message-ID so that Mailman will set its own Message-ID and we can actually send messages to more than one list.
I'm 100% aware that this is not a Mailman issue. If I'm too far OT, I send my apologies.
Alex Collins. Library Systems and Support Officer.
Rivermead Library. Tel: 0845 XXX XXXX
alex.collins (at) anglia.ac.uk http://libweb.anglia.ac.uk
This product is printed with 100% recycled electrons !
EMERGING EXCELLENCE: In the Research Assessment Exercise (RAE) 2008,
more than 30% of our submissions were rated as 'Internationally
Excellent' or 'World-leading'. Among the academic disciplines now rated
'World-leading' are Allied Health Professions & Studies; Art & Design;
English Language & Literature; Geography & Environmental Studies;
History; Music; Psychology; and Social Work & Social Policy &
Administration. Visit www.anglia.ac.uk/rae for more information.
This e-mail and any attachments are intended for the above named
recipient(s)only and may be privileged. If they have come to you in
error you must take no action based on them, nor must you copy or show
them to anyone please reply to this e-mail to highlight the error and
then immediately delete the e-mail from your system. Any opinions
expressed are solely those of the author and do not necessarily
represent the views or opinions of Anglia Ruskin University. Although
measures have been taken to ensure that this e-mail and attachments are
free from any virus we advise that, in keeping with good computing
practice, the recipient should ensure they are actually virus free.
Please note that this message has been sent over public networks which
may not be a 100% secure communications
Email has been scanned for viruses by Altman Technologies' email
management service - www.altman.co.uk/emailsystems
I am relatively new to Mailman. I am administering a server with around
200 lists on it. I am setting up Swish-e to provide searchable list
archives. Each list has it's own search. In the search form, I need to
be able to reference a cgi script.
Each list has its own cgi script in the mailman cgi-bin folder. The
names of the cgi scripts are the same as the folder containing the
mailing list in mailman.
I thought I could just use %(listname)s in my templates to reference
this. Unfortunately, this is returning the friendly list name, not the
name of the folder where the list resides. I can change my cgi script
names if I need to, but I was wondering if there is a template tag I can
use to insert the name of the list as it appears in the mailman lists
Also, is there a list of the mailman tags that can be used in these