Page MenuHomePhabricator

OAbot crashes and requires a restart
Closed, ResolvedPublic

Description

Recently I've had to restart OAbot a few times after it stopped responding, usually in the middle of intense usage. I see a pattern here:

Tue May 28 10:29:26 2019 - uwsgi_response_sendfile_do(): Broken pipe [core/writer.c line 636] during GET /oabot/static/css/bootstrap.min.css (192.168.180.0)
IOError: write error
SIGINT/SIGQUIT received...killing workers...
--
Mon Jul 22 21:02:39 2019 - SIGPIPE: writing to a closed pipe/socket/fd (probably the client disconnected) on request /oabot/process?name=Eric_Kandel (ip 192.168.180.0) !!!
SIGINT/SIGQUIT received...killing workers...
--
mapped 363840 bytes (355 KB) for 4 cores
*** Operational MODE: preforking ***
mounting /data/project/oabot/www/python/src/app.py on /oabot
WSGI app 0 (mountpoint='/oabot') ready in 34 seconds on interpreter 0xfd9490 pid: 1 (default app)
*** uWSGI is running in multiple interpreter mode ***                                                                                                                                                     spawned uWSGI master process (pid: 1)
spawned uWSGI worker 1 (pid: 10, cores: 1)                                                                                                                                                                spawned uWSGI worker 2 (pid: 11, cores: 1)
spawned uWSGI worker 3 (pid: 12, cores: 1)
spawned uWSGI worker 4 (pid: 13, cores: 1)
SIGINT/SIGQUIT received...killing workers...
--
SIGINT/SIGQUIT received...killing workers...
--
Mon Sep  2 08:32:55 2019 - uwsgi_response_sendfile_do(): Connection reset by peer [core/writer.c line 636] during GET /oabot/static/css/bootstrap.min.css (192.168.180.0)
IOError: write error
SIGINT/SIGQUIT received...killing workers...
--
SIGINT/SIGQUIT received...killing workers...                                                                                                                                                              --
Fri Sep  6 09:51:10 2019 - SIGPIPE: writing to a closed pipe/socket/fd (probably the client disconnected) on request /oabot/static/css/bootstrap.min.css (ip 192.168.180.0) !!!                           Fri Sep  6 09:51:10 2019 - uwsgi_response_sendfile_do(): Broken pipe [core/writer.c line 636] during GET /oabot/static/css/bootstrap.min.css (192.168.180.0)
IOError: write error
SIGINT/SIGQUIT received...killing workers...                                                                                                                                                              --
SIGINT/SIGQUIT received...killing workers...

Mon Sep 16 10:26:40 2019 - uwsgi_response_write_body_do(): Broken pipe [core/writer.c line 322] during GET
IOError: write error
SIGINT/SIGQUIT received...killing workers...                                                                                                                                                              --
SIGINT/SIGQUIT received...killing workers...                                                                                                                                                              --

Fri Sep 20 14:57:59 2019 - uwsgi_response_write_body_do(): Broken pipe
IOError: write error
SIGINT/SIGQUIT received...killing workers...

Event Timeline

Nemo_bis triaged this task as Medium priority.Sep 20 2019, 4:09 PM
Nemo_bis created this task.

Meanwhile for mitigation: https://github.com/dissemin/oabot/pull/57

I suspect this has to do with some users who may have a slow connection.

Nemo_bis claimed this task.

The errors in the task description happened even after the fix for T216170. This has not happened in a while, although there was some usage. I'm tentatively considering it resolved.