Commit Graph

17 Commits

Author SHA1 Message Date
Vishnu Ks 1b0b135bfc management: Remove unused imports from process_queue. 2017-08-08 14:13:19 -07:00
Umair Khan 0e8231d0f1 process_queue: Recover gracefully after PostgreSQL restart.
- For threaded workers:
Django's autoreloader catches SIGQUIT(3) to reload the program. If
a process being watched by autoreloader exits with status code 3,
reloader will restart the process. To reload, we send SIGUSR1(10)
signal from consumers to a handler in process_queue which then
exits with status code 3.

- For single worker per process:
Catch the SIGUSR1 and quit; supervisorctl will restart the worker
automatically.

Fixes #5512
2017-07-07 16:33:15 -07:00
rht 940cf9db3b Run queue processors multithreaded in production if system memory <3.5GB.
While running queue processors multithreaded will limit the
performance available to very small systems, it's easy to fix that by
adding more RAM, and previously, Zulip didn't work on such systems at
all, so this is unambiguously an improvement there.

Fixes #32.
Fixes #34.

(Commit message expanded significantly by tabbott.)
2017-06-03 12:19:58 -07:00
Rafid Aslam 41bd88d5ed pep8: Fix E301 pep8 violations.
Fix "E301: expected (1 or 2) blank line" pep8 violations.
2016-11-29 08:51:44 -08:00
Steve Howell ca43cbc654 logging: Reducing logging in run-dev.py for queue workers.
In dev, we no longer log that individual queue workers were launched.

Instead, in dev (and prod as well), we log a message with the total
count.
2016-11-17 11:12:02 -08:00
Tim Abbott 86e933a4a1 process_queue: Suppress USING_RABBITMQ warnings in test suite. 2016-10-27 12:36:06 -07:00
Tim Abbott 1e54897ca7 process_queue: Add missing type annotation. 2016-08-04 15:57:03 -07:00
Tim Abbott 92062b9526 process_queue: Add missing annotation. 2016-08-04 15:57:02 -07:00
rahuldeve a3745178e5 Use django.utils.autoreload to restart queue workers at code change.
Fixes #621, #1045.
2016-06-26 20:12:11 -07:00
Tim Abbott a1a27b1789 Annotate most Zulip management commands. 2016-06-04 10:12:06 -07:00
Tim Abbott 06b33da709 process_queue: Fix missing worker.setup() in single-threaded codepath. 2016-03-27 23:17:16 -07:00
Tim Abbott cd2348e9ae Run queue processers multithreaded in development.
This change drops the memory used for Python processes run by Zulip in
development from about 1GB to 300MB on my laptop.

On the front of safety, http://pika.readthedocs.org/en/latest/faq.html
explains "Pika does not have any notion of threading in the code. If
you want to use Pika with threading, make sure you have a Pika
connection per thread, created in that thread. It is not safe to share
one Pika connection across threads.".  Since this code only connects
to rabbitmq inside the individual threads, I believe this should be
safe.

Progress towards #32.
2016-03-20 18:04:24 -07:00
Tim Abbott 7595e4b05f process_queue: Fix worker variable being accessed before initialization. 2016-02-03 19:29:44 -08:00
Reid Barton ae0ae3dde8 Django 1.8: declare positional arguments in management commands
(imported from commit d9efca1376de92c8187d25f546c79fece8d2d8c6)
2015-08-20 23:35:40 -07:00
Zev Benjamin ffb6266319 Refuse to run a queue processor if USING_RABBITMQ is False
(imported from commit 39beff47cdbb18ba39756989e6f07facbd16864f)
2013-10-28 14:30:53 -04:00
Zev Benjamin 871afde142 Prepare process_queue to be used with supervisor's numprocs
(imported from commit 5a652b93f1e8b32b5ed89d622035161abaedfb11)
2013-09-24 20:44:15 -04:00
Zev Benjamin 6cf5ebc63d Add new system for defining and running queue-processing workers
(imported from commit 7e77da57b8ad0b70837785c85f546601ba5b1957)
2013-09-24 20:44:15 -04:00