September 2022 - Mailman-users - lists.mailman3.org

Re: mail stuck in "out" queue
by Dmitry Makovey 16 Sep '22

16 Sep '22

So our mailman got stuck again, but I did follow your advise. Here's what I've found: On 3/29/19 2:58 PM, Mark Sapiro wrote: > On 3/29/19 2:28 PM, Dmitry Makovey wrote: >> I've had encountered this situation now twice: mail is being stuck in >> "out" queue and nothing in the logs indicates the issue. Is there a way >> to ask mailman-core politely to attempt delivery (perhaps with increased >> debugging in the logs) *without* restarting the core (which does solve >> *this* problem but creates other problems, like availability of the >> mailman as a whole) > > > Is the "out" runner running? yes: # docker top mailman-core | grep -F out: 100 21681 13215 0 Mar28 ? 01:14:16 /usr/local/bin/python /usr/local/bin/runner --runner=out:0:1 -C /etc/mailman.cfg > ps -fwwA |grep runner=out > > If not, you should be able to start it as the Mailman user with a > command like > > /path/to/python3 /path/to/mailman/bin/runner -C /path/to/mailman.cfg > --runner=out:0:1 > > See 'ps -fwwA|grep runner' for examples > > If the out runner is running, is there a file named *.bak along with all > the *.pck files in queue/out/ there is one, yes: # ls -l /srv/mail/mailman/core/var/queue/out/*.bak -rw-rw---- 1 100 65533 21512 Apr 14 19:45 /srv/mail/mailman/core/var/queue/out/1555296331.8386078+2c905825b82708e5a682b872b0aff320c9dd0b2e.bak what does it indicate and what should my steps be to resolve this (other than restarting queues)? > And if it's not, is there anything in Mailman's logs to indicate why it > died? > > Also, 'mailman restart' should only take a matter of seconds, but if the > out runner is running but not processing its queue, I think sending it a > SIGINT should restart it. SIGINT killed the process allright, but it didn't auto-respawn. I ended up spawning it with: /usr/local/bin/python /usr/local/bin/runner --runner=out:0:1 -C /etc/mailman.cfg but soon realized it doesn't go to background nicely so I've nohup'ed it in the end. but it did solve my problem. -- Sr System and DevOps Engineer SoM IRT

5 12

Held message notifications ending up in junk
by Tibor Molnar 15 Sep '22

15 Sep '22

Hi there, I wonder if anyone else if having this trouble that all the held message notifications are landing in junk folder. Regards, Tibor

4 5

Messages not going to held queue
by Mail Person 14 Sep '22

14 Sep '22

Hi all, Not sure why, but messages set to be held for moderation are appearing to be discarded. I've tried creating a new test list, adjusting settings at the user and list level, but when a user sends a message that should be held for moderation, nothing happens. smtp.log shows the message is going through but it doesn't get to the moderation queue.

3 8

Message-Footer after migrating to Mailman3
by christopher.claus＠tgcamberg1848.de 13 Sep '22

13 Sep '22

Hello together, I migrated a number of lists from mailman2 to mailman3 using the import functionality. Everthing works fine and the new lists are in use. Currently I´m using the maxking-docker-images with postorius: Mailman Core Version GNU Mailman 3.3.3 (Tom Sawyer) Mailman Core API Version 3.1 Mailman Core Python Version 3.8.5 (default, Feb 19 2021, 09:27:16) [GCC 9.3.0] After this migration I started a refactoring with the list-styles specially with message-footers. Already the old lists in mailman2 were confused in some configuration details an I was clear to import this "misconfiguration" and clean it up afterwards. But this was much more difficult than expected: some lists are using a footer (from the mailman2-config) and there is no record in the template-section visible. Further I found no configuration item which points to this "old" footer. I found the file [code] /usr/lib/python3.8/site-packages/mailman/templates/de/list:member:generic:footer.txt [/code] which includes the old footer: [code] _______________________________________________ $display_name Mailingliste -- $listname Zur Abmeldung von dieser Mailingliste senden Sie eine Nachricht an ${short_listname}-leave@${domain} [/code] but no configuration item to disable this behaviour. I created a new template list:member:regular:footer for the list to override the footer. But the old footer is still used - in detail: the resulting message sent to all subcribers of the list contains only the footer and the content sent to the list is an attachement of that "footer-message". This is the basic reason why I want to change it. In the logfiles I found an error and a warning: [code] ERROR 2021-05-24 16:22:22,895 1648 django.security.DisallowedHost Invalid HTTP_HOST header: 'tgc_mailman_web:8000'. The domain name provided is not valid according to RFC 1034/1035. WARNING 2021-05-24 16:22:22,896 1648 django.request Bad Request: /postorius/api/templates/list/homepage.gruppe.tgcamberg1848.de/list:member:regular:footer [/code] Obviously I have a problem with the hostname. I set [code] POSTORIUS_TEMPLATE_BASE_URL=http://tgc_mailman_web:8000 [/code] where tgc_mailman_web is my container name. This URL ist stored in the DB-table "template" for the name=list:member:regular:footer. A call to my fqdn or to localhost [code] wget http://localhost:8000/postorius/api/templates/list/homepage.gruppe.tgcamber… [/code] will return the footer itself - this works fine. Therefore, should i set POSTORIUS_TEMPLATE_BASE_URL to http://localhost:8000? I think, resolving this issue will use the newly configured footer. But I started with the intention to remove alle footers due to the attachement situation described above. Does anybody have an idea how to change the configuration of my list? Are these problems probably related to the import? Currently we are using more then 25 lists with a lot of users. Manually recreating all lists is not possible an could only be a test for a single list. Best regards, chrclaus

6 18

"The content of this message was lost."
by skenny＠scss.tcd.ie 12 Sep '22

12 Sep '22

I see that this issue has been discussed in a few other postings - it seems to be affecting us here at the moment. We're running Mailman 3.3.5 inside maxking's mailman-core/mailman-web Docker containers (version 0.4.3). This morning I composed a mail message and included two of our mailing lists in the To: field. On the first list I'm a member with address of the form my_addr_1(a)this.com. On the second I'm on it with a different address, my_addr_2(a)that.com. In the case of each list, Mailman held the message for moderation. As administrator, I accepted both. I then got the full message in my inbox at my_addr_1(a)this.com, but in my other inbox at my_addr_2(a)that.com, the body of the message was not present. Instead it read The content of this message was lost. It was probably cross-posted to multiple lists and previously handled on another list. I'm wondering what's the best thing to do here. Should I maybe downgrade to an earlier version of Mailman (i.e. an earlier version of maxking's Docker containers), pending resolution of this issue ...? Thanks for any advice. Stephen PS: The Mailman web interface "System Information" link reports: Mailman Core Version GNU Mailman 3.3.5 (Tom Sawyer) Mailman Core API Version 3.1 Mailman Core Python Version 3.8.10 (default, May 6 2021, 06:30:44) [GCC 9.3.0]

2 2

How to set uWSGI buffer size in mailman3 with Dcoker
by bob B 08 Sep '22

08 Sep '22

Running into an issue where users are getting a 502 bade gateway and we traced it down to the headers that are getting sent when they request web pages. This looks like our issue ********************* By default, uWSGI allocates a very small buffer (4096 bytes) for the headers of each request. If you start receiving “invalid request block size” in your logs, it could mean you need a bigger buffer. Increase it (up to 65535) with the buffer-size option. ********* So any idea on how to set this in mailman3 running in the maxking docker environment?

2 2

mailman3 not working for mailman21 lists
by blup＠fortytwo.ch 06 Sep '22

06 Sep '22

I installed mailman3 on a server where mailman21 was installed before. Newly created mailinglists work fine. But if I create mailinglists with a name that was previously used on mailman21 nothing works (same behaviour with and without import of old configs and archives; no logs). I purged all old mailman21 config and data and also old postfix aliases (or at least I do think so). If I send a mail to old list list1(a)a.b and new list list2(a)a.b I get the following logs /var/log/mail.log Aug 21 22:17:40 basilisgg postfix/lmtp[1728]: ED2171120BF: to=<list2(a)c.a.b>, orig_to=<list2(a)a.b>, relay=127.0.0.1[127.0.0.1]:8024, delay=0.31, delays=0.17/0.01/0/0.13, dsn=2.0.0, status=sent (250 Ok) Aug 21 22:17:40 basilisgg postfix/lmtp[1728]: ED2171120BF: to=<list1(a)c.a.b>, orig_to=<list1(a)a.b>, relay=127.0.0.1[127.0.0.1]:8024, delay=0.31, delays=0.17/0.01/0/0.13, dsn=2.0.0, status=sent (250 Ok) /var/log/mailman3/smtp.log ug 21 22:17:40 2022 (934) Available AUTH mechanisms: LOGIN(builtin) PLAIN(builtin) Aug 21 22:17:40 2022 (934) Peer: ('127.0.0.1', 44746) Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) handling connection Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) Data: b'LHLO c.a.b' Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) Data: b'MAIL FROM:<....>' Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) sender: .... Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) Data: b'RCPT TO:<list2(a)c.a.b>' Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) recip: list2(a)c.a.b Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) Data: b'RCPT TO:<list1(a)c.a.b>' Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) recip: list1(a)c.a.b Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) Data: b'DATA' Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) Data: b'QUIT' Aug 21 22:17:40 2022 (934) ('127.0.0.1', 44746) connection lost Aug 21 22:17:40 2022 (934) Connection lost during _handle_client() Aug 21 22:17:46 2022 (936) <166111306097.933.1775779871266718007@...> smtp to list2(a)a.b for 1 recips, completed in 3.1437485218048096 seconds How can I get more info on why the mail for list1 isn't processed? Where could old mailman21 or postfix info on list1 be stored that causes mailman3 to not handle the list properly? How can I diff the config between list1 and list2 from the command line? Regards Diana

5 12