Add documentation about requiring high port numbers for gunicorn about _old_vespene HOT 5 CLOSED

Hi @muki,

As it is intended to work, 0_common.sh is just the input to all the other scripts.

You would have to rerun 6_services.sh after changing the port.

When you say the app dies, I'm a bit curious how that happened because nothing would know about that change?

Did you mean you ran some of the other scripts?

Logs go into /var/log/vespene/ - the supervisor log probably contains any failures about starting the service.

As such, I'm inclined to thing this isn't a bug as written, but maybe something we need to clarify in the docs.

Am I reading that correctly?

Thanks!

from _old_vespene.

So I changed the port to 1000 in the 0_common.sh before running anything else. I read through it and decided I want to proxy vespene through 1000. This was my first edit to the file. I then went through and read all the other setup scripts and changed nothing else. I ran them in order from 0 to 6. After I was done I set up nginx and when I tried to open it up, NGINX threw a 502 bad gateway error. This made me believe the app itself somehow crashed. These are the logs from that time:

supervisord.log

2018-11-01 21:12:23,200 CRIT could not write pidfile /var/run/supervisord.pid
2018-11-01 21:12:24,204 INFO spawned: 'worker_tutorial-pool0' with pid 10730
2018-11-01 21:12:24,207 INFO spawned: 'server' with pid 10731
2018-11-01 21:12:25,670 INFO success: worker_tutorial-pool0 entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2018-11-01 21:12:25,670 INFO success: server entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)
2018-11-01 21:12:30,039 INFO exited: server (exit status 1; not expected)

web.log

[2018-11-01 21:12:24 +0000] [10731] [INFO] Starting gunicorn 19.9.0
[2018-11-01 21:12:24 +0000] [10731] [ERROR] Retrying in 1 second.
[2018-11-01 21:12:25 +0000] [10731] [ERROR] Retrying in 1 second.
[2018-11-01 21:12:26 +0000] [10731] [ERROR] Retrying in 1 second.
[2018-11-01 21:12:27 +0000] [10731] [ERROR] Retrying in 1 second.
[2018-11-01 21:12:28 +0000] [10731] [ERROR] Retrying in 1 second.
[2018-11-01 21:12:29 +0000] [10731] [ERROR] Can't connect to ('127.0.0.1', 1000)

worker_tutorial-pool.log

2018-11-01 21:12:25,669 - vespene - INFO - serving queue: tutorial-pool

I tried restarting the service and nothing changed. After I went back to 0_common.sh and changed the gunicorn line to port 8000 I reran 6_services.sh vespene started to work. This led me to believe that this was a bug, and that you can't really run vespene on a loopback with port 1000.

from _old_vespene.

Oh, yeah, so that's it.

Vespene runs as the 'vespene' user, so if you are just running it like that you need to be above port 1024.

https://stackoverflow.com/questions/10182798/why-are-ports-below-1024-privileged

If that is indeed it - and it seems to be - it seems like gunicorn is not throwing a good error about not being able to bind to a privileged port.

From the error messages here - https://stackoverflow.com/questions/40121221/gunicorn-not-starting-on-port-80-but-starting-on-port-8000 - that seems to be exactly it.

It seems the best answer here is to comment that it needs a port > 1024 in 0_common.sh

(You could of course technically run it as root, but that's something you should really want to avoid anyway)

I'll leave this open so I can document this better.

Thank you very much for the heads up!

from _old_vespene.

Somehow I didn't know about the 1024 limit. Today I learned ...

from _old_vespene.

Closing per #73 - thanks for the report!

from _old_vespene.