I have the following displayed when I look at the "System Information"
Mailman Core Version GNU Mailman 3.3.0 (La Villa Strangiato)
Mailman Core API Version 3.0
Mailman Core Python Version 3.7.6 (default, Dec 19 2019, 23:49:42)
[GCC 5.4.0 20160609]
What surprises me is the protocol version. I have tried to set the
protocol version to 3.1 in mailman core (which should be the latest
according to the documentation) and restarting both core and web but
that gives no effect. Normally I have not had any protocol version
defined. Everything seems to work just fine but something is strange,
either the documentation, my setup or mailman/postorius.
I have struggled for hours but cannot seem to figure it out. After
registering new user accounts in Postorius, the "confirm your address"
emails don't get sent via SMTP (my MTA does not even log connection
attempts), but their content is displayed in uwsgi-error.log.
I think I disabled developer mode in Core and Postorius:
$ mailman conf | grep devmode
[devmode] enabled: no
[devmode] testing: no
[devmode] wait: 60s
$ grep ^DEBUG settings.py
DEBUG = False
Users can subscribe via list-join(a)domain.tld, including confirmations.
Account registration using the WebUI works as well, it's just the emails
that don't get sent. What am I missing, I wonder?
Forgive me if this is the wrong list.
I’m curious to hear what users are doing for MM3 hosting and what you are using for outgoing services.
I am hosting a Mailman3 installation for a community that I am part of, and we have 6 lists which generated 136,000 outgoing emails last month. I’m currently hosting this on a 2-core 4GB RAM EC2 (t3a.medium) instance at Amazon. I went with Amazon because their outgoing email relay (SES) is free for the first 62,000 emails/month. (I didn’t do my homework - I thought that would be plenty!) I’d much rather be hosting on Linode, since their setup is waaaaay simpler and easier to understand.
I went with an SMTP service so that I wouldn’t have to deal with managing an IP address reputation…. (Any other SMTP relay service besides the Amazon SES/hosting would cost >$80/month.)
Do you use an SMTP relay service?
Do you just send directly from your VM’s IP address?
Do you have trouble with getting black listed?
What have you done about your IP address reputation?
Is DKIM and SPF enough to get past an IP address without an established reputation?
WE are currently using mailman2. I have the task to setup and install a maiman3 server. Postfix is installed and working. Can anybody point me to a link where the actual installation is described using RHEL 7 or RHEL 8. This will be a hardened DISA compliant server, so I am trying to avoid installing compilers and tools that are not part of the standard RHEL 7 repos.
I am struggling with my mailman server for quite some time now. I am on
3.3 and the out runner cannot sent emails any more. According to this
issue  I found out--by logging the exception as well--, that mailman
has this problem:
Mar 31 09:34:51 2020 (71057) ACCEPT: <c1e4aa1be5b61c289c69a66a6d55565333d5a35b.camel(a)rumpelsepp.org>
Mar 31 09:34:52 2020 (71079) Cannot connect to SMTP server 127.0.0.1 on port 25: [Errno 21] Is a directory: '/usr/lib/python3.8/site-packages/mailman/templates'
What the hell is going on here? ☹︎ any ideas what's wrong?
The runner doesn't try to connect to postfix, since it fails with a
strange template error.
Whilst we are all stuck in doors I have started a project to move Mailman3 from a Docker installation to a new server using a Python venv. Old server was Debian Jessie and the new one is Buster. Using Exim with stock config.
Everything seems to be going ok but mail isn't delivering (cannot connect to SMTP server).
I have /opt/mailman which I did chown -R list:list
mailman.cfg has following MTA settings:
I wrote a Systemd unit which starts Mailman from the venv and started up Mailman once I got Postgres sorted so it would create the necessary files in the var directory structure.
One thing I noticed and had to fix is that it doesn't creat the directories for existing lists, and the Exim config examples out there check for the existance of those directories to identify whether the address exists. I created those anyway and set ownership perms as before.
Once I fixed that I sent a test message to the owner address and that routed correctly and the headers indicated the message was sent back to localhost by Mailman.
However I can't get it to route a list message, here is a log showing the start of the Mailman master processes and me sending a test message from Outlook (so incoming mail is routing ok):
Mar 28 16:03:20 2020 (21877) Master started
Mar 28 16:03:32 2020 (21889) lmtp runner started.
Mar 28 16:03:32 2020 (21896) virgin runner started.
Mar 28 16:03:32 2020 (21887) command runner started.
Mar 28 16:03:32 2020 (21890) nntp runner started.
Mar 28 16:03:32 2020 (21886) bounces runner started.
Mar 28 16:03:33 2020 (21892) pipeline runner started.
Mar 28 16:03:33 2020 (21888) in runner started.
Mar 28 16:03:33 2020 (21895) retry runner started.
Mar 28 16:03:34 2020 (21894) rest runner started.
[2020-03-28 16:03:34 +0000]  [INFO] Starting gunicorn 20.0.4
[2020-03-28 16:03:34 +0000]  [INFO] Listening at: http://127.0.0.1:8001 (21894)
[2020-03-28 16:03:34 +0000]  [INFO] Using worker: sync
[2020-03-28 16:03:34 +0000]  [INFO] Booting worker with pid: 21963
Mar 28 16:03:34 2020 (21885) archive runner started.
[2020-03-28 16:03:34 +0000]  [INFO] Booting worker with pid: 21965
Mar 28 16:03:34 2020 (21897) digest runner started.
Mar 28 16:03:36 2020 (21891) out runner started.
Mar 28 16:04:25 2020 (21888) ACCEPT:
Mar 28 16:04:26 2020 (21885) HyperKitty archived message <AM0PR08MB39085771D0E2D9114FDE8E83AFCD0(a)AM0PR08MB3908.eurprd08.prod.outlook.com> to https://email@example.com…
[28/Mar/2020:16:04:26 +0000] "GET /3.0/lists/blind-sysadmins(a)lists.hodgsonfamily.org HTTP/1.1" 200 445 "-" "GNU Mailman REST client v3.3.0"
Mar 28 16:04:27 2020 (21891) Cannot connect to SMTP server localhost on port 25
[28/Mar/2020:16:04:27 +0000] "GET /3.0/lists/blind-sysadmins(a)lists.hodgsonfamily.org/config HTTP/1.1" 200 2595 "-" "GNU Mailman REST client v3.3.0"
[28/Mar/2020:16:04:27 +0000] "GET /3.0/users/andrew(a)hodgson.io HTTP/1.1" 200 439 "-" "GNU Mailman REST client v3.3.0"
Now in the out queue I have this:
root@elrond:/opt/mailman# ls -la /opt/mailman/var/queue/out
drwxrwx--- 2 list list 4096 Mar 28 16:34 .
drwxr-xr-x 14 list list 4096 Mar 28 13:25 ..
-rw-rw---- 1 list list 11152 Mar 28 16:34 1585413255.2412584+390ab148e589108ed9e849445e5e90410f69def3.bak
-rw-rw---- 1 list list 11152 Mar 28 16:34 1585413255.4891446+75bde47331619263b05f582c28ceb85a13bb5265.pck.tmp
There is no connection attempt in the Exim log from Mailman and I can connect to localhost:25.
This is a dualstack machine and I was wondering whether we had any IPV6 issues (I had an issue with Gunicorn and Nginx routing traffic due to this earlier that I had to sort), but as owner addresses are working I don't know what is different on that.
My new MM3 server is receiving these errors from cox.net:
"(delivery temporarily suspended: host cxr.mx.a.cloudfilter.net[220.127.116.11] refused to talk to me: 421 cxr-ibgw-6004a.stratus.cloudmark.com cmsmtp too many sessions from 18.104.22.168. Refer to Error Codes section at https://www.cox.com/residential/support/email-error-codes.html for more information. AUP#CXCNCT)
The link gives just this:
421 - [too many sessions from <source IP>] The sending IP address has exceeded
the five maximum concurrent connection limit. Configure your sending server to establish
no more than five concurrent connections.
1. Is that a legitimate excuse to refuse passing on emails or is cox.net too stingy?
2. Is that a setting we need to change in Mailman somewhere? Or should this be handled in some Postfix setting?
I have a few (but not all) members complaining that their digests are full of coded characters. Two examples below were forwarded to me:
> What I ended up doing was creating 2 "practice lines" in the centre of=20
> our big teaching circle (I always teach new figures in a big circle) -=20
> which each had 3 people set up in a line with generous space between,=20
> and first=C2=A0 I walked through one of those lines myself starting first=
> Robins pass Right=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Larks Loop L=
What is the cause? Is there anything I can do about it?
Now that we are on mailman3, I'm finding things that were really useful in
mailman2 are completely missing in mailman3, which seems odd, since they in
particular affect the ability of list moderators to do their job well.
a) When a post was made to a moderated list in mailman2, I would get an
email that included a LINK to the held email. Now I just get a canned
email saying I need to go to the dashboard, blah blah. Why was this link
removed? It was extremely useful and now I have to jump through extra
b) When a post is made to a moderated list in mailman2, when I went to the
page to approve/deny/discard, there was additionally an option to remove
the moderated flag from the user for all future emails. This option no
longer exists. This was EXTREMELY useful, as we force all new members of
the list to be moderated for a time while we evaluate the quality of their
postings. Now I have to go manually track down the user and their
settings. Unnecessary extra hoops to jump through
c) When a post is made to a moderated list, one of the options when
selecting REJECT was it allowed me to set a custom message when rejecting
the email. This was extremely useful. For example today, someone emailed
our bug tracking list directly instead of filing a bug. In the past, I
would have added a note to them about what they should do (in this case, I
would have told them to send their email to our community technical
discussion list). Now I can no longer do that, I can only reject it or
discard it. Why was this removed?
Packaged, certified, and supported LDAP solutions powered by OpenLDAP: