Search results for "Postorius" - Mailman-users

Cannot setup HTTPS support for mailman3-web

by jean-christophe manciot

Ubuntu 20.04 mailman3 3.2.2-1 mailman3-web 0+20180916-10 python3-django 2:2.2.9-2ubuntu1 In ```/etc/mailman3/uwsgi.ini``` (example.org is defined in the local /etc/hosts as 10.0.0.1): ``` [uwsgi] # Port on which uwsgi will be listening. ### suwsgi: uwsgi protocol over TLS uwsgi-socket = example.org:8430 # https://uwsgi-docs.readthedocs.io/en/latest/HTTPS.html?highlight=ssl https = /etc/ssl/example.org.crt,/etc/ssl/example.org_rsakey.pem.decrypted,HIGH ... # Drop privileges and don't run as root. uid = www-data gid = www-data ... ``` I suppose that uwsgi reads the certificate & its private key before dropping root privileges. Right now, uwsgi cannot access them as ```www-data```. In ```/etc/mailman3/mailman-web.py``` (symlinked to django settings.py) : ``` ... SECURE_PROXY_SSL_HEADER = ('HTTP_X_FORWARDED_PROTO', 'https') SECURE_PROXY_SSL_HEADER = ('HTTP_X_FORWARDED_SCHEME', 'https') # Other security settings # SECURE_SSL_REDIRECT = True # If you set SECURE_SSL_REDIRECT to True, make sure the SECURE_REDIRECT_EXEMPT # contains at least this line: # SECURE_REDIRECT_EXEMPT = [ # "archives/api/mailman/.*", # Request from Mailman. # ] # SESSION_COOKIE_SECURE = True # SECURE_CONTENT_TYPE_NOSNIFF = True # SECURE_BROWSER_XSS_FILTER = True # CSRF_COOKIE_SECURE = True # CSRF_COOKIE_HTTPONLY = True # X_FRAME_OPTIONS = 'DENY' ... POSTORIUS_TEMPLATE_BASE_URL = 'https://example.org:8430' ``` In ```/etc/nginx/sites-enabled/example.org.conf``` ``` ... location / { include /etc/nginx/uwsgi_params; # upstream app server uwsgi_pass suwsgi://example.org:8430; } ``` uwsgi is listening on the right address/port: ``` # networks-list-connections.sh | grep :8430 tcp 0 0 10.0.0.1:8430 0.0.0.0:* LISTEN 33 54336045 1768187/uwsgi ``` Despite those settings and all the relevant systemd services restarted, I get ```502 Bad Gateway``` while browsing to ```https://example.org/postorius/lists/?all-lists``` (or any other page) with the following error nginx message: ``` *1 peer closed connection in SSL handshake while SSL handshaking to upstream, client: 10.0.0.1, server: example.org, request: "GET /postorius/lists/?all-lists HTTP/2.0", upstream: "suwsgi://10.0.0.1:8430", host: "example.org", referrer: "https://example.org/postorius/lists/" ``` Of course, without those specific https/suwsgi settings, everything works fine. What am I missing/doing incorrectly? Does uwsgi need access to the certificate/private key while running as non-root for instance?

6 years, 9 months

Any possible way to reduce robot spam accounts?

by Stephen J. Turnbull

Franklin Weng writes: > I really have no idea how the robots find my system since I haven't > told anyone the ip or URL of my mailman instance.� And, they're > registering on the system, not the mailing list since I have only > one list for testing now and it wasn't assaulted by spam yet. The bots just check every domain on the internet for the /mailman3 and /postorius paths. > * just let users join from email (lists-join(a)lists.url) instead of > registering on the mailman instance?� or What I do is open up the firewall only to ips I have legit users at. You probably can't do that, but if you can, it's quite effective. You could restrict it to email-only subscription in one of the following ways, but the bots have known how to do that since the 90s, and the only control you have is to require moderator approval. Requiring address verification doesn't help much, the bots have been doing that almost as long. You can disable the administrative UI entirely, by removing the entries for mailman3 and postorius from the top-level urls.py. You can edit the template for the list information to remove the subscription form. > * use a scheme like CAPTCHA to reduce the successful opportunity of > robot registering? I think somebody has implemented CAPTCHA for Mailman 3. Search the archives of this list. But the bots also know how to solve CAPTCHAs, and for some CAPTCHAs they were faster and more accurate than humans. Unless your users expect instant membership, or you expect an awful lot of subscriptions, I would just let subscriptions accumulate for a day or two and then discard them all. You can also write a script to figure out the IPs and ban those hosts at the firewall. Steve

3 years, 3 months

Re: Guidance Request: Migrating Existing Mailman 3 Installation to Docker-Based Deployment

by Prasanth Nair

Hi Odhiambo and Stephen, Thank you both for your helpful responses and guidance. Given that this is a critical production system, I am now considering migrating to a new server using a host-based Mailman installation first, before evaluating a Docker-based setup at a later stage. In that case, could you please advise on the recommended approach for migrating from my current host-based Mailman 3 installation to a new host-based server? In particular, I would appreciate any guidance on: - Safely migrating the PostgreSQL databases (e.g., dump and restore approach) - Preserving mailing lists, memberships, and configurations - Migrating HyperKitty archives and Postorius data - Any best practices to ensure a smooth cutover with minimal downtime Any documentation or pointers you can share would be greatly appreciated. Thank you again for your time and support. Kind regards, Prasanth Nair On Sat, 11 Apr 2026 at 17:52, Stephen J. Turnbull <steve(a)turnbull.jp> wrote: > Prasanth Nair writes: > > > I am currently running a Mailman 3 setup on a standalone server and > > planning to migrate it to a Docker-based deployment using the official > > Docker Mailman images. I would appreciate your guidance on the > recommended > > approach for this migration. > > docker-mailman is basically a hobby of the maintainer's. It's not > even maintained in the GitLab project. As far as I know it's there > because it's frequently requested and folks who run docker usually > know what they're doing so it doesn't impose a particular burden on > us. But really the only expert on docker-mailman is Abhilash himself. > > All three of the major Mailman suite applications (core, Postorius, > and HyperKitty) use object-relation managers with automated schema > updates. So contrary to Odhiambo's advice, I would try to replicate > the versions of the Mailman components and PostgreSQL, then do > upgrades in-place in Docker. Doing a database dump (in the current > host) and restore (in the containerized PostgreSQL) is safe if you > have the same schema on both hosts. (Upgrading the PostgreSQL version > is much less risky but why tempt the software gremlins?) > > An alternative if you know PostgreSQL well would be to setup a > "failover and failback" system, where you start with your current > PostgreSQL as the primary and the container's as a replica, then when > you're ready to go live you swap primary and replica for the database > and point the DNS to the containerized Mailman. This probably > involves less downtime (although in my experience PostgreSQL dump and > restore is quite fast, and the Mailman databases for core and > Postorius are small), but the big advantage is if you have a problem > with the containers you can switch back right away. I know this works > because I worked on a project where high availability was required, > but another guy did the PostgreSQL work so I can't provide details. > > Networking is a FAQ for mailman-docker. I've only run it out of the > box with all its defaults, and that worked fine. But somehow people > have problems with it. My guess is that they're trying to fit the > Mailman containers into some predetermined scheme, either a Docker > host that hosts other containers and there are IP conflicts, or some > organizational constraints. If you know what you're doing, no problem > I think, but if you're newish to containers, you might want to set up > docker on a host dedicated to Mailman. > > Other than that, I think Abhilash provided a preferred channel for > discussing docker-mailman in the docs in his repo, that would be where > to get your best (and more detailed) information. > > -- > GNU Mailman consultant (installation, migration, customization) > Sirius Open Source https://www.siriusopensource.com/ > Software systems consulting in Europe, North America, and Japan > _______________________________________________ > Mailman-users mailing list -- mailman-users(a)mailman3.org > To unsubscribe send an email to mailman-users-leave(a)mailman3.org > https://lists.mailman3.org/mailman3/lists/mailman-users.mailman3.org/ > Archived at: > https://lists.mailman3.org/archives/list/mailman-users@mailman3.org/message… > > This message sent to prasanth.nair(a)linaro.org >

6 months, 3 weeks

Re: Attachement processings questions

by Mark Sapiro

On 2/3/20 4:27 AM, Lars Schimmer wrote: > > In a mailinglist it seems some mails are not forwarded to a member, but > is listed in archive in hyperkitty. > That mail did include a PDF attachement, but HyperKitty does not list > any attachement. There are two issues here. Why wasn't the message delivered to list members?, and why is there no attachment in Hyperkitty? For the first issue, you need to examine mailman.log for the time of the post looking for errors and tracebacks. Also, the MTA logs may have useful info if the messages were sent but not delivered. For the second, there are many content filtering settings not (yet) exposed in Postorius. These include filter_extensions, filter_types, pass_extensions and pass_types which control which MIME parts are removed. For a newly created Mailman 3 list, these four settings are empty, but importing a Mailman 2.1 list will set them according to the corresponding MM 2.1 list's settings. > All I found was "convert HTML to plaintext", but does this setting > remove attachements? It would only affect HTML parts, not pdfs. > Is there any generel documentation about attachement processing in MM3? <https://mailman.readthedocs.io/en/latest/src/mailman/handlers/docs/filterin…>. > I would like this mailinglist to accept mails with attachements, send it > to the members and archvie the attachments in HyperKitty, to. You can set Settings -> Alter Messages -> Filter content to No in Postorius to not do any content filtering at all. This will override the Collapse alternatives and Convert html to plaintext settings and pass the message without removing any MIME parts. But, I think you must have other issues, either with this mail in particular or in general as evidenced by the mail not being received by list members. -- Mark Sapiro <mark(a)msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan

6 years, 9 months

Re: Problem with languages

by Abhilash Raj

> On Mar 20, 2021, at 10:19 AM, Stephen J. Turnbull <turnbull.stephen.fw(a)u.tsukuba.ac.jp> wrote: > > Andreas Barth writes: > >> Well, I personally wouldn't be so afraid, but that might come from the >> fact that I know the debian ecosystem way too well :) > > FWIW I've run a testing + a very few pinned to sid[1] for 15 years now, > and never had unexpected problems with testing packages themselves, > and only very rarely with distro issues (mucked up dependencies where > I couldn't install the current version of some package as specified in > testing). > > J. Guenther writes: > >>>> Also, translations had been the reason for some discussions lately, so >>>> this is still some work-in-progress. (Also on the git version I'm >>>> using, there are more texts in English than I'd wish for.) >>> >>> What I hear is that this is a common issue and I probably should get >>> along with it for now!? > > As Andi says, there was a glitch in the Weblate (IIRC) to Mailman > sources pipeline which should be fixed now, but for specific languages > you kind of have to track the completion rate with that translation > community. Probably there's a way to check on Weblate itself. Yeah, you can go the project page on Weblate and check completion rate for a specific language. For example this is the page for Postorius. https://hosted.weblate.org/projects/gnu-mailman/postorius/ > If > Weblate seems to be much farther along than Mailman git, ping us, it's > still a pretty manual process. For web interface (Postorius, django-mailman3 and Hyperkitty) it is only manual to the point that someone has to click “Merge” ;-) The process is fairly automated and new translations are sent as Merge Request[1] against our Git repos nightly. Since each commit adds a Merge Commit, I sometimes wait a little for 4-5 commits to pile up before merging them. But if there are large number of people installing from Git for up-to-date translations, I wouldn’t mind merging them asap. We definitely merge everything before a release is made and translations in a released version will definitely lag behind the Weblate for any changes made after the release. Core has a bit of manual work needed to copy translations from .po files to actual template files (.txt), but hopefully I or someone else will be able to fix that soon. It just needs a script to read from all the language specific .po files and write the translated templates to individual .txt files (where Core expects them to be). [1]: Like this one, for example, https://gitlab.com/mailman/postorius/-/merge_requests/618 > > Regards, > Steve > > Footnotes: > [1] These are all applications that I had extensive experience with > install from source, and still want quite recent versions. > > > _______________________________________________ > Mailman-users mailing list -- mailman-users(a)mailman3.org > To unsubscribe send an email to mailman-users-leave(a)mailman3.org > https://lists.mailman3.org/mailman3/lists/mailman-users.mailman3.org/ -- thanks, Abhilash Raj (maxking)

5 years, 7 months

Re: old usernames appearing in mailman.log

by Mark Sapiro

On 6/25/23 10:57 PM, Ken Alker wrote: > > I have studied the output, looked at the code lines referenced in the > Traceback, and I can't figure out the problem (I feel like the sad guy > on the side of the street with his hood up staring into the engine bay > with stream pouring out and not knowing what to do next). I did some > analyses on the log file and determined that there are 71 messages it is > having problems processing and there are 71 messages "stuck" in > /opt/mailman/mm/var/archives/hyperkitty/spool/ (and that it has tried to > process each message about 20 times). Each time there is a new post, it retries all the messages in the spool/ directory. > Since upgrading to V3.3.8 I've > not seen new messages get stuck like this (although it's only been > running smoothly for 24 hours now and there hasn't been much activity). > But I'm suspecting there was something awry between the conversion from > V2 to V3 in February and me installing V3.3.8 that caused this (based on > message time stamps). It is HIGHLY likely these messages are in the > archiving spool as a result of me unshunting the shunt queue and that > they are copies of the messages that got delivered to the subscribers > when I did that... but are now hung up in the archiving process. Are new messages being archived OK. If so, that is a real mystery. It says mailman-hyperkitty can talk to hyperkitty and archive new messages, but it can't talk to hyperkitty when attempting to process a queued message. > Following is an excerpt from mailman.log. Note that the "first section" > (from DOCTYPE at top to 404 at bottom) is repeated four times (maybe not > identical; but similar) (so I did not include it to save room) before > the Traceback appears. > > Jun 25 00:13:04 2023 (386086) No cached copy of the public suffix list > found > Jun 25 00:13:04 2023 (386086) ACCEPT: > <7500f0e2-1ae1-309d-60d7-6e592c7abf7a(a)west.net> > Jun 25 00:13:04 2023 (386090) HyperKitty failure on > https://lists.netlojix.com/hyperkitty/api/mailman/urls: I just went to that URL and got a 404. This is a serious problem. https://lists.netlojix.com/hyperkitty/api/ also returns a 404. It should return some documentation of the API. https://lists.netlojix.com/archives/api/ does work and returns the expected information. Your urls.py file probably contains in its urlpatterns list ``` re_path(r'^mailman3/', include('postorius.urls')), re_path(r'^archives/', include('hyperkitty.urls')), ``` The simplest fix is to add ``` re_path(r'^postorius/', include('postorius.urls')), re_path(r'^hyperkitty/', include('hyperkitty.urls')), ``` to that list. Another way to fix this is in your mailman-hyperkitty.cfg, you have ``` base_url: https://lists.netlojix.com/hyperkitty ``` Change that to ``` base_url: https://lists.netlojix.com/archives ``` but it wouldn't hurt to add the `^postorius/` and `^hyperkitty/` patterns to urls.py anyway. -- Mark Sapiro <mark(a)msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan

3 years, 4 months

help for new installation.

by claudio.jofre＠yandex.com

Hello everyone. A few days ago, I am trying to install and configure mailman 3, in a Centos7. and the truth has not been achieved. I am already a little desperate. I'll tell you First I install the hyperkitty-el.repo repository and install mailman 3, hyperkitty and postorius. Initially when reviewing and making some minor modifications to the mailman configuration file, the service apparently rises without major problems. The problem occurs when I try to configure the web interfaces. Initially the apache server did not start, with a series of errors. Following the instructions of the document published here: http://www.admin-magazine.com/Archive/2015/29/Mailman-3.0-a-long-time-comin… I managed to eliminate the initial errors, which prevented to raise the Apache server. However, when trying to access the web pages, in the case of hyperkitty, I get an error 500 with an "internal server error". In the case of postorius, I get an error 403 forbidden. In the mail, I receive a log, with the following content: Traceback (most recent call last): File "/ usr / bin / django-admin", line 5, in <module> management.execute_from_command_line () File "/usr/lib/python2.7/site-packages/django/core/management/__init__.py", line 354, in execute_from_command_line utility.execute () File "/usr/lib/python2.7/site-packages/django/core/management/__init__.py", line 303, in execute settings.INSTALLED_APPS File "/usr/lib/python2.7/site-packages/django/conf/__init__.py", line 48, in __getattr__ self._setup (name) File "/usr/lib/python2.7/site-packages/django/conf/__init__.py", line 44, in _setup self._wrapped = Settings (settings_module) File "/usr/lib/python2.7/site-packages/django/conf/__init__.py", line 92, in __init__ mod = importlib.import_module (self.SETTINGS_MODULE) File "/usr/lib64/python2.7/importlib/__init__.py", line 37, in import_module __import __ (name) ImportError: No module named settings In the Apache logs, it also shows the following: [Mon Dec 11 21: 48: 41.875069 2017] [authz_core: error] [pid 15333] [client my-client-ip: 33584] AH01630: client denied by server configuration: / etc / postorius / sites / default / srv when I execute the command: django-admin check throws me the following: django.core.exceptions.ImproperlyConfigured: Requested setting DATABASES, but settings are not configured. You must either define the environment variable DJANGO_SETTINGS_MODULE or call settings.configure() before accessing settings. checking the ports, I can see that python has open ports 8000 in 127.0.0.1, and 8001 in external ip, for more than I have looked for information, documentation, etc, about the procedure of the installation. all the documentation I found, part of the premise of an installation, clean from git, by pip, but none from the yum and rpm. So it is difficult for me to determine what to apply, and what not. apparently no one has installed, or has had problems in the installation from repositories, or I am omitting some step (most likely) that I could not determine. the bad thing, is that in postfix, I have everything configured, with digital signatures, so I would not like to lose all that work, as to start with an installation from 0. If you can help me, it would be very helpful. since I made the request of help in another list, of dev, to which it is supposed that the consultations should be made, according to the historical mails, but I never received answer. Of course, thank you very much for any help and guidance they can give me.

8 years, 10 months

Filter response from Mailman3 held messages queue.

by Stephen J. Turnbull

Andy Matthews writes: > I've found some libraries which can parse the message property and > extract the relevant bits, but I'd rather pass some sort of flag or > filter to the initial Mailman 3 API call and get back less > information. Before I express my opinion, note that I am not authoritative on this, but I have a decent record on channeling the consensus. So YMMV FWIW I'd be against putting such a thing into the core without a lot of beta testing, I think. I imagine there would be various opinions on how much less and which data should come back. There is a plugin architecture which specifically can be used to implement REST routes: https://docs.mailman3.org/projects/mailman/en/latest/src/mailman/plugins/do… While doing it this way undoubtedly makes it more complicated for the initial implementer, I think it would be a lot simpler (more self-contained) for people who want to implement different filters, etc. > I've looked through the documentation but I'm not finding anything > of the sort. I don't think there is for this, although there's a certain amount of code for filtering the messages that are being processed (stripping attachments of specified MIME types, converting text/html to text/plain, etc). I'm not sure if that's done in advance of holding for moderation. So maybe it would be OK to repurpose that code to your end (though I'd want it to be a proper refactoring, not cargo culting the post distribution filter into the moderation code), and somebody would need to do the same for Postorius. Not necessarily you (I infer you're talking about a separate application that talks to Mailman core for moderation), but availability to Postorius users with common administrative UI would make it a lot more attractive to us.

1 year, 9 months

Re: Apache+mod_wsgi issue

by Odhiambo Washington

On Mon, Dec 26, 2022 at 6:35 PM Guillermo Hernandez (Oldno7) via Mailman-users <mailman-users(a)mailman3.org> wrote: > El 26/12/22 a las 16:08, Odhiambo Washington escribió: > > On Mon, Dec 26, 2022 at 5:07 PM Eggert Ehmke<eggert(a)eehmke.de> wrote: > > Odhiambo, I didn't change anything from my config as I wrote down in > https://forums.FreeBSD.org/threads/mailman-3.61050/post-488128, save the > python version (I upgraded to 3.9 moths ago, as I posted in this list). > > My urls.py looks like this: > > ################### > from django.conf import settings > ### mod. 30/10/22 > ###from django.conf.urls import include, url > from django.conf.urls import include > from django.urls import re_path as url > ##### > > from django.contrib import admin > from django.urls import reverse_lazy > from django.views.generic import RedirectView > > urlpatterns = [ > url(r'^$', RedirectView.as_view( > url=reverse_lazy('list_index'), > permanent=True)), > url(r'^postorius/', include('postorius.urls')), > url(r'^hyperkitty/', include('hyperkitty.urls')), > url(r'', include('django_mailman3.urls')), > url(r'^accounts/', include('allauth.urls')), > # Django admin > url(r'^admin/', admin.site.urls), > url(r'^mailman3/', include('postorius.urls')), > url(r'^archives/', include('hyperkitty.urls')), > ] > > Hi @Guillermo Hernandez (Oldno7) <guillermo(a)querysoft.es> Does your https://site.name/user-profile/ work? PS: I deviated from installing things on the base system and went the virtualenv way for a production server. I believe it's a lot easier to maintain as compared to having everything using FreeBSD ports/packages. Oh, and virtualenv is well-supported. Remember FreeBSD as on OS is not even supported, which has led to most FreeBSD admins staying with MM2 or going the way of Sympa and whatever else:) -- Best regards, Odhiambo WASHINGTON, Nairobi,KE +254 7 3200 0004/+254 7 2274 3223 "Oh, the cruft.", egrep -v '^$|^.*#' ¯\_(ツ)_/¯ :-)

3 years, 10 months

Re: 120 gmail users removed from lists !

by Yvan ♞

Hello, Thanks for this answer, my problem is I'm on yunohost server and the update 3.3.9 is not available now. Any "handmaid" solution should be great. Sincerely Le 22/10/2023 à 10:45, Stephen J. Turnbull a écrit : > elemac(a)free.fr writes: > > > Hello, on 15th October, 120 users with gmail address where suddenly > > removed from lists in mailman3, something changed in google ? > > Google is evil, who knew? > > Google decided to publish a DMARC policy of "p=none". This announces > to the world that you have to decide for yourself whether to accept > mail from Gmail with a broken DKIM signature from gmail.com. Gmail > has decided that they will not accept such email from themselves, even > though they know more about the sender than anyone else, and even have > access to the original sender copy. > > Mark has a patch that applies your preferred DMARC mitigation to Gmail > authors unconditionally. I think it's in the new version just > released 3.3.9, so you can upgrade. The name of the variable is > dmarc_addresses, and in Postorius it will be in the same settings > group as other DMARC settings, with a similar tag. > > If for some reason you strongly prefer not to upgrade immediately but > want this feature, it's commit #8532f751b6b8bfee3210c1327d755756f99d99ac > (you can shorten that to 8532f). I do not recommend this. It probably > is independent of other changes in 3.3.9 from other releases. But if > not you could have issues. > > I think you also need the patches in Postorius, but I don't think > they've been released yet. "Coming soon" as they say. The relevant > commit is #770add1ebf27fd335c675607f26bb8f7cf313ddd, > > If you search back in the archives, I think Mark posted a link to his > original patches. > > Steve >

3 years

Mailman-users search results for query "Postorius"