Web Installer preventing NodeBB server from starting.

[timc]

After filling out the Web Installer form and then running ./nodebb start on the command line. I get the errors listed below.
It seems like the Web Installer is still using the 4567 port the node server is trying to use.
I have tried killing all node processes before running ./nodebb start but the processes refuse to be killed and this has no effect.

Any suggestions?

./nodebb start errors

Starting NodeBB
  "./nodebb stop" to stop the NodeBB server
  "./nodebb log" to view server output
  "./nodebb help" for more commands


NodeBB v2.8.5 Copyright (C) 2013-2023 NodeBB Inc.
This program comes with ABSOLUTELY NO WARRANTY.
This is free software, and you are welcome to redistribute it under certain conditions.
For the full license, please visit: http://www.gnu.org/copyleft/gpl.html

Clustering enabled: Spinning up 1 process(es).

2023-01-31T18:55:54.079Z [4567/192] - info: Initializing NodeBB v2.8.5 https://node.mydomain.com
2023-01-31T18:55:56.476Z [4567/192] - info: [socket.io] Restricting access to origin: https://node.mydomain.com:*
(node:192) [LRU_CACHE_UNBOUNDED] UnboundedCacheWarning: TTL caching without ttlAutopurge, max, or maxSize can result in unbounded memory consumption.
(Use `node --trace-warnings ...` to show where the warning was created)

2023-01-31T18:55:56.727Z [4567/192] - warn: [plugins/load] The following plugins may not be compatible with your version of NodeBB. This may cause unintended behaviour or crashing. In the event of an unresponsive NodeBB caused by this plugin, run `./nodebb reset -p PLUGINNAME` to disable it.
  * nodebb-rewards-essentials

2023-01-31T18:55:56.837Z [4567/192] - info: [api] Adding 0 route(s) to `api/v3/plugins`
2023-01-31T18:55:56.853Z [4567/192] - info: [router] Routes added
2023-01-31T18:55:56.865Z [4567/192] - info: 🎉 NodeBB Ready
2023-01-31T18:55:56.867Z [4567/192] - info: 🤝 Enabling 'trust proxy'
2023-01-31T18:55:56.870Z [4567/192] - error: NodeBB address in use, exiting...
Error: listen EADDRINUSE: address already in use 0.0.0.0:4567
    at Server.setupListenHandle [as _listen2] (node:net:1733:16)
    at listenInCluster (node:net:1781:12)
    at doListen (node:net:1930:7)
    at process.processTicksAndRejections (node:internal/process/task_queues:83:21)
2023-01-31T18:55:56.874Z [4567/192] - error: uncaughtException: listen EADDRINUSE: address already in use 0.0.0.0:4567
Error: listen EADDRINUSE: address already in use 0.0.0.0:4567
    at Server.setupListenHandle [as _listen2] (node:net:1733:16)
    at listenInCluster (node:net:1781:12)
    at doListen (node:net:1930:7)
    at process.processTicksAndRejections (node:internal/process/task_queues:83:21) {"date":"Tue Jan 31 2023 18:55:56 GMT+0000 (Coordinated Universal Time)","error":{"address":"0.0.0.0","code":"EADDRINUSE","errno":-98,"port":4567,"syscall":"listen"},"exception":true,"os":{"loadavg":[0.81,0.68,0.78],"uptime":603592.54},"process":{"argv":["/usr/local/bin/node","/usr/src/app/app.js"],"cwd":"/usr/src/app","execPath":"/usr/local/bin/node","gid":1000,"memoryUsage":{"arrayBuffers":869872,"external":2935100,"heapTotal":110329856,"heapUsed":78592824,"rss":174796800},"pid":192,"uid":1000,"version":"v18.13.0"},"stack":"Error: listen EADDRINUSE: address already in use 0.0.0.0:4567\n    at Server.setupListenHandle [as _listen2] (node:net:1733:16)\n    at listenInCluster (node:net:1781:12)\n    at doListen (node:net:1930:7)\n    at process.processTicksAndRejections (node:internal/process/task_queues:83:21)","trace":[{"column":16,"file":"node:net","function":"Server.setupListenHandle [as _listen2]","line":1733,"method":"setupListenHandle [as _listen2]","native":false},{"column":12,"file":"node:net","function":"listenInCluster","line":1781,"method":null,"native":false},{"column":7,"file":"node:net","function":"doListen","line":1930,"method":null,"native":false},{"column":21,"file":"node:internal/process/task_queues","function":"process.processTicksAndRejections","line":83,"method":"processTicksAndRejections","native":false}]}
2023-01-31T18:55:56.875Z [4567/192] - error: Error: listen EADDRINUSE: address already in use 0.0.0.0:4567
    at Server.setupListenHandle [as _listen2] (node:net:1733:16)
    at listenInCluster (node:net:1781:12)
    at doListen (node:net:1930:7)
    at process.processTicksAndRejections (node:internal/process/task_queues:83:21)
2023-01-31T18:55:56.875Z [4567/192] - info: [app] Shutdown (SIGTERM/SIGINT) Initialised.
2023-01-31T18:55:56.878Z [4567/192] - error: Error [ERR_SERVER_NOT_RUNNING]: Server is not running.
    at new NodeError (node:internal/errors:400:5)
    at Server.close (node:net:2075:12)
    at Object.onceWrapper (node:events:627:28)
    at Server.emit (node:events:525:35)
    at emitCloseNT (node:net:2128:8)
    at process.processTicksAndRejections (node:internal/process/task_queues:81:21)
[cluster] Child Process (192) has exited (code: 1, signal: null)
[cluster] Spinning up another process...

----- snip -----
# Repeats the above 2 more times

3 restarts in 10 seconds, most likely an error on startup. Halting.

[timc]

I figured out a solution. I just needed to prevent the web installer from starting up and claiming the 0.0.0.0:4567 address the nodebb server tries to use.

The Web Installer only starts if there isn't already a config.json file. So I added a ConfigMap to my kubernetes manifest to insert a valid config.json file into the correct directory when the nodebb pod starts. I used this post for guidance in setting up the ConfigMap.

Thanks for your help!

[phenomlab]

@timc do you have any systemd or perhaps pm2 daemons active?

[timc]

I don't think so.

node@nodebb-694ff7946-56wtz:/usr/src/app$ systemctl --version
bash: systemctl: command not found
node@nodebb-694ff7946-56wtz:/usr/src/app$ pm2 --version
bash: pm2: command not found

[phenomlab]

@timc if you use the cli installer, you can use another port. However, I'm curious - do you have an existing installation of NodeBB running in this specific machine?

[julian]

Might be worth it to run ps aux | grep node to see if NodeBB is indeed still running

[timc]

@phenomlab I am trying to deploy nodebb in kubernetes using the official docker container. So I don't think I can change the port with the cli installer.

@julian my ps aux | grep node output is below. There does appear to be a nodeBB server running, I think it is the web installer? However, all my attempts to kill the processes (kill -9 <PID>, etc.) seem to have no effect (the processes restart).

node@nodebb-694ff7946-56wtz:/usr/src/app$ ps aux | grep node
node           1  0.0  0.0   2484   364 ?        Ss   Jan31   0:00 /bin/sh -c test -n "${SETUP}" && ./nodebb setup || node ./nodebb build; node ./nodebb start
node           8  0.0  3.2 42845668 128480 ?     Sl   Jan31   0:15 node ./nodebb build
node          48  0.2  0.0   4164  3180 pts/0    Ss   18:15   0:00 bash
node          56  0.0  0.0   6760  2748 pts/0    R+   18:15   0:00 ps aux
node          57  0.0  0.0   3244   704 pts/0    S+   18:15   0:00 grep node

[PitaJ]

@timc have you tried killall node?

[timc]

@PitaJ unfortunately, killall is not installed on the kubernetes pod (Debian 11), and since pods aren't supposed to have root access I cannot install it.

[timc]

I figured out a solution. I just needed to prevent the web installer from starting up and claiming the 0.0.0.0:4567 address the nodebb server tries to use.

The Web Installer only starts if there isn't already a config.json file. So I added a ConfigMap to my kubernetes manifest to insert a valid config.json file into the correct directory when the nodebb pod starts. I used this post for guidance in setting up the ConfigMap.

Thanks for your help!

[julian]

Great! Thanks for checking in with your solution