Message shows up individually but not in the thread
Hi,
I've just observed something weird. While some users were testing the new mailman setup, the reply did not make it into the corresponding thread, but is still associated with it. That is, the message does not show up in the thread page at1. However, the reply is visible if requested individually2. Now what's weird is that in 2 there's a link at the top that says "< thread" that takes one to 1, so there is some kind of association appearently made but only in one direction (the reverse one). mailman.log has this to say:
Nov 03 17:22:31 2017 (498) ACCEPT: <1fa6f9e1-254c-e999-42d0-9a6cc263020d@openmailbox.org>
Nov 03 17:22:32 2017 (507) 127.0.0.1 - - "GET /3.0/users/sydneyjd@openmailbox.org HTTP/1.1" 200 305
Nov 03 17:22:32 2017 (495) HyperKitty archived message <1fa6f9e1-254c-e999-42d0-9a6cc263020d@openmailbox.org> to https://lists.secretchronicles.org/hyperkitty/list/tsc-users@lists.secretchronicles.org/message/FDANAZ3ED4R2XRARPEZSFJWZCWTWSIUY/The mailman-suite log contains:
INFO 2017-11-03 17:22:32,910 14330 hyperkitty.views.mailman Archived message <1fa6f9e1-254c-e999-42d0-9a6cc263020d@openmailbox.org> to https://lists.secretchronicles.org/hyperkitty/list/tsc-users@lists.secretchronicles.org/message/FDANAZ3ED4R2XRARPEZSFJWZCWTWSIUY/This looks to me like a bug. The OP of that thread sent the same message to our other mailinglist3 (where he did not get a reply), so Mailman may has been confused by that.
Marvin
-- Blog: https://www.guelkerdev.de PGP/GPG ID: F1D8799FBCC8BC4F
On 11/03/2017 11:36 AM, Marvin Gülker wrote:
I've just observed something weird. While some users were testing the new mailman setup, the reply did not make it into the corresponding thread, but is still associated with it.
Are you running all of Django's periodic tasks?
See <http://docs.mailman3.org/en/latest/config-web.html#scheduled-tasks-required>
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
Am 03. November 2017 um 12:35 Uhr -0700 schrieb Mark Sapiro <mark@msapiro.net>:
Are you running all of Django's periodic tasks?
They are running. This is the excerpt from /etc/crontab:
# Mailman jobs
@daily mailman bash -c 'cd /home/mailman && . /home/mailman/env3/bin/activate && mailman digests --send'
@hourly mailman bash -c 'cd /home/mailman && . /home/mailman/env2/bin/activate && django-admin runjobs hourly --pythonpath /home/mailman/mailman-suite/mailman-suite_project --settings settings'
@daily mailman bash -c 'cd /home/mailman && . /home/mailman/env2/bin/activate && django-admin runjobs daily --pythonpath /home/mailman/mailman-suite/mailman-suite_project --settings settings'
@weekly mailman bash -c 'cd /home/mailman && . /home/mailman/env2/bin/activate && django-admin runjobs weekly --pythonpath /home/mailman/mailman-suite/mailman-suite_project --settings settings'
@monthly mailman bash -c 'cd /home/mailman && . /home/mailman/env2/bin/activate && django-admin runjobs monthly --pythonpath /home/mailman/mailman-suite/mailman-suite_project --settings settings'
@yearly mailman bash -c 'cd /home/mailman && . /home/mailman/env2/bin/activate && django-admin runjobs yearly --pythonpath /home/mailman/mailman-suite/mailman-suite_project --settings settings'
* * * * * mailman bash -c 'cd /home/mailman && . /home/mailman/env2/bin/activate && django-admin runjobs minutely --pythonpath /home/mailman/mailman-suite/mailman-suite_project --settings settings'
2,17,32,47 * * * * mailman bash -c 'cd /home/mailman && . /home/mailman/env2/bin/activate && django-admin runjobs quarter_hourly --pythonpath /home/mailman/mailman-suite/mailman-suite_project --settings settingsObviously, mailman-suite is installed at /home/mailman/mailman-suite and owned by user 'mailman', with the actual Django project in the subdirectory 'mailman-suite_project', just as cloned from the Git repository. The only difference from the official guide here is that I took the liberty to place Mailman into two virtualenvs (a Python2 one for the Django project, and a Python3 one for Mailman-Core*).
There are no errors reported with this in /var/log/cron.log. And before you ask, the task queue is running as well. The uWSGI config has this:
# Setup the django_q related worker processes.
attach-daemon = ./manage.py qclusterAnd the qcluster process shows both up in the process list and is happily logging away in the uWSGI log.
Greetings Marvin
- It'd be great if the mailman-suite was moved to Python 3 at some point. That way one wouldn't need two separate virtualenvs.
-- Blog: https://www.guelkerdev.de PGP/GPG ID: F1D8799FBCC8BC4F
On 11/04/2017 01:19 AM, Marvin Gülker wrote:
Am 03. November 2017 um 12:35 Uhr -0700 schrieb Mark Sapiro <mark@msapiro.net>:
Are you running all of Django's periodic tasks?
They are running. This is the excerpt from /etc/crontab: ... Obviously, mailman-suite is installed at /home/mailman/mailman-suite and owned by user 'mailman', with the actual Django project in the subdirectory 'mailman-suite_project', just as cloned from the Git repository. The only difference from the official guide here is that I took the liberty to place Mailman into two virtualenvs (a Python2 one for the Django project, and a Python3 one for Mailman-Core*).
There are no errors reported with this in /var/log/cron.log. And before you ask, the task queue is running as well. The uWSGI config has this:
# Setup the django_q related worker processes. attach-daemon = ./manage.py qclusterAnd the qcluster process shows both up in the process list and is happily logging away in the uWSGI log.
There is a possibly related issues at <https://gitlab.com/mailman/hyperkitty/issues/151> and some older ones at <https://gitlab.com/mailman/hyperkitty/issues/131> and <https://gitlab.com/mailman/hyperkitty/issues/149>.
I'm sorry I can't offer more help than that at this point.
- It'd be great if the mailman-suite was moved to Python 3 at some point. That way one wouldn't need two separate virtualenvs.
We're working on moving Postorius and HyperKitty to MM 3.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
Am 04. November 2017 um 10:51 Uhr -0700 schrieb Mark Sapiro <mark@msapiro.net>:
There is a possibly related issues at <https://gitlab.com/mailman/hyperkitty/issues/151> and some older ones at <https://gitlab.com/mailman/hyperkitty/issues/131> and <https://gitlab.com/mailman/hyperkitty/issues/149>.
I'm sorry I can't offer more help than that at this point.
Okay. I'll have an eye on the problem and see whether I can make out any pattern on messages not appearing. I'm not (yet?) seeing the other symptoms described in #151, but our mailinglists are usually rather quiet until some burst of discussion arises. There have not been any newer posts than the linked ones yet.
The mails all went out properly via email. It's a problem related to Hyperkitty thus.
Marvin
-- Blog: https://www.guelkerdev.de PGP/GPG ID: F1D8799FBCC8BC4F
Am 04. November 2017 um 20:34 Uhr +0100 schrieb Marvin Gülker <m-guelker@phoenixmail.de>:
I'm not (yet?) seeing the other symptoms described in #151, but our mailinglists are usually rather quiet until some burst of discussion arises. There have not been any newer posts than the linked ones yet.
Don't name it or it happens. This new thread post1 never reached the front page. Can Hyperkitty be forced somehow to rebuild its entire internal structure, ideally by reading all mails in again?
Marvin
-- Blog: https://www.guelkerdev.de PGP/GPG ID: F1D8799FBCC8BC4F
Am 04. November 2017 um 21:52 Uhr +0100 schrieb Marvin Gülker <m-guelker@phoenixmail.de>:
Don't name it or it happens. This new thread post[1] never reached the front page. Can Hyperkitty be forced somehow to rebuild its entire internal structure, ideally by reading all mails in again?
Sorry for spamming. It *just now* reached the front page. I was too impacient.
Sorry for the noise! I'd still be interested in a forced-rebuild command, though.
Marvin
-- Blog: https://www.guelkerdev.de PGP/GPG ID: F1D8799FBCC8BC4F
On 11/04/2017 01:52 PM, Marvin Gülker wrote:
Don't name it or it happens. This new thread post[1] never reached the front page. Can Hyperkitty be forced somehow to rebuild its entire internal structure, ideally by reading all mails in again?
django-admin update_index
(It can take a while)
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
Am 04. November 2017 um 14:02 Uhr -0700 schrieb Mark Sapiro <mark@msapiro.net>:
On 11/04/2017 01:52 PM, Marvin Gülker wrote:
Don't name it or it happens. This new thread post[1] never reached the front page. Can Hyperkitty be forced somehow to rebuild its entire internal structure, ideally by reading all mails in again?
django-admin update_index
(It can take a while)
Thanks!
-- Blog: https://www.guelkerdev.de PGP/GPG ID: F1D8799FBCC8BC4F
participants (2)
-
Mark Sapiro -
Marvin Gülker