Docker compose "web" service fails to reload on code changes

[yeraydiazdiaz]

Recently I've been noticing the reloading of the web service errors forcing me to restart the docker compose service. Has anyone else experienced this?

Hupper enters a cycle where it can't bind to the port because it is already in use. Here are the logs:

web_1            | [2019-09-04 09:31:05 +0000] [16] [INFO] Worker exiting (pid: 16)
web_1            | Server did not exit, forcefully killing.
web_1            | Starting monitor for PID 17.
web_1            | {"logger": "raven.base.Client", "level": "INFO", "event": "Raven is not configured (logging is disabled). Please see the documentation for more information.", "thread": 140079147017984}
web_1            | [2019-09-04 09:31:31 +0000] [17] [INFO] Starting gunicorn 19.9.0
web_1            | [2019-09-04 09:31:31 +0000] [17] [ERROR] Connection in use: ('0.0.0.0', 8000)
web_1            | [2019-09-04 09:31:31 +0000] [17] [ERROR] Retrying in 1 second.
web_1            | [2019-09-04 09:31:32 +0000] [17] [ERROR] Connection in use: ('0.0.0.0', 8000)
web_1            | [2019-09-04 09:31:32 +0000] [17] [ERROR] Retrying in 1 second.
web_1            | [2019-09-04 09:31:33 +0000] [17] [ERROR] Connection in use: ('0.0.0.0', 8000)
web_1            | [2019-09-04 09:31:33 +0000] [17] [ERROR] Retrying in 1 second.
web_1            | [2019-09-04 09:31:34 +0000] [17] [ERROR] Connection in use: ('0.0.0.0', 8000)
web_1            | [2019-09-04 09:31:34 +0000] [17] [ERROR] Retrying in 1 second.
web_1            | [2019-09-04 09:31:35 +0000] [17] [ERROR] Connection in use: ('0.0.0.0', 8000)
web_1            | [2019-09-04 09:31:35 +0000] [17] [ERROR] Retrying in 1 second.
web_1            | [2019-09-04 09:31:36 +0000] [17] [ERROR] Can't connect to ('0.0.0.0', 8000)
web_1            | Press ENTER or change a file to reload.
web_1            | Traceback (most recent call last):
web_1            |   File "<string>", line 1, in <module>
web_1            |   File "/opt/warehouse/lib/python3.7/site-packages/hupper/ipc.py", line 318, in spawn_main
web_1            |     func(**kwargs)
web_1            |   File "/opt/warehouse/lib/python3.7/site-packages/hupper/worker.py", line 263, in worker_main
web_1            |     func(*spec_args, **spec_kwargs)
web_1            |   File "/opt/warehouse/lib/python3.7/site-packages/hupper/reloader.py", line 285, in wait_main
web_1            |     input('')
web_1            | EOFError: EOF when reading a line
web_1            | Starting monitor for PID 30.

... repeats endlessly...

[di]

I've noticed this too. Thanks for filing an issue about it!

[yeraydiazdiaz]

This is happening again but the output is different:

web_1            | /opt/warehouse/src/warehouse/search/queries.py changed; reloading ...
web_1            | Gracefully killing the server.
web_1            | [2020-02-26 20:51:14 +0000] [9] [INFO] Handling signal: term
web_1            | [2020-02-26 20:51:14 +0000] [16] [INFO] Worker exiting (pid: 16)
web_1            | Server did not exit, forcefully killing.
web_1            | Starting monitor for PID 17.
web_1            | {"logger": "raven.base.Client", "level": "INFO", "event": "Raven is not configured (logging is disabled). Please see the documentation for more information.", "thread": 139630094440192}
web_1            | [2020-02-26 20:51:32 +0000] [17] [INFO] Starting gunicorn 20.0.4
web_1            | [2020-02-26 20:51:32 +0000] [17] [ERROR] Connection in use: ('0.0.0.0', 8000)                                                                                                                                              web_1            | [2020-02-26 20:51:32 +0000] [17] [ERROR] Retrying in 1 second.
web_1            | [2020-02-26 20:51:33 +0000] [17] [ERROR] Connection in use: ('0.0.0.0', 8000)
web_1            | [2020-02-26 20:51:33 +0000] [17] [ERROR] Retrying in 1 second.

[di]

I thought #7404 might resolve this (the changelog looks promising) but I'm still seeing this as well.

[di]

I'm also still experiencing this with versions of hupper that used to work.

[di]

@yeraydiazdiaz Looks like hupper made a new release which might solve this: #7479

[di]

The new hupper release did not seem to solve this. Downgrading to gunicorn 19.x did not resolve it either.