Technical Support

This likely cropped up because flag actions now use the v3 API instead of socket.io, so it could be it overloaded the backend, which started sending 503s.

yeah, thank you. I've solved it with deleting the database. In the logs I found the error of skipping the creation of the admin user, because it already exists.

@marksibly I would recommend doing it through the reverse proxy (nginx caddy or whatever you use), and not to mediate through the node.js
Caddyfile example (https://caddyserver.com) :

Then set the address in the config.json file of the nodebb:

and then run

And the forum will be available at https://mywebsite.com/forum/ (be sure to keep the / at the end)

@baris Thanks a lot for replying, and I apologize for wasting your time 😞
I somehow got the markdown plugin disabled while upgrading, hence the issue.
Enabling it back up solved it.

@Thmhe said in Running the forum after restarting the computer:

@josef said in Running the forum after restarting the computer:

nodebb.bat

?What I need to write in the program is

nodebb.bat

How should I proceed from here?
b22a02f7-4f82-4f3a-8616-4e66ece020a5-image.png

@josef said in Running the forum after restarting the computer:

and add a "start" argument

@baris Thanks - that's a good point actually. I might have some custom CSS or something in the headers that points to that. I'll check.

EDIT: That was it. Removed that folder and subfolders, and we're good 🙂

@julian said in 503 Errror on Production, OAuth2 callback not working:

which in Docker might be exposed somewhere...

Learn how to write to, view, and configure a container's logs

Docker Documentation (docs.docker.com)

Yes thank you

What version of NodeBB?

You should see the relevant error in the output log. You should be able to capture the message by running ./nodebb log and then trying the test email.

Have you tried the same SMTP settings in a separate email client?

@baris Perfect. Thanks very much for clarifying!

@julian Thanks much, I'll give that a try!

can we do slug translation now? (since it has been many updates)

You can experiment with some other guides and projects.

As you probably seen before:

(hub.docker.com)

One more option is to make your own modified Dockerfile. For example Nobbic works on Alpine Linux based Docker images and most of the NodeBB guides are Ubuntu/Debian based. So first thing is to check the host machine/hosting provider and what is going on when you execute Dockerfile. Including Docker version which was mentioned many times on this forum.

After you link the plugin so its in node_modules, you need to activate it either from the ACP or from the CLI with ./nodebb activate nodebb-plugin-my-plugin

Topics on the sitemap are filtered here https://github.com/NodeBB/NodeBB/blob/master/src/sitemap.js#L132, unless you gave guests the View deleted posts privilege they shouldn't show up in the sitemap.

The sitemap is also cached so it would take some time for the entries to disappear from the sitemap.

It's still but if I changed a theme it will be Okay.

That field is passed to https://github.com/expressjs/cors#configuration-options, you might be able to just leave it out If not specified, defaults to reflecting the headers specified in the request's Access-Control-Request-Headers header.

Did you try changing your url field in config.json to http://127.0.0.1:4567? And change the connection code to also use 127.0.0.1:4567 instead of localhost.

@PitaJ Ty!

Thank you both and apologies for the delay!

I will try them both out.

@o-o - I was not able to make fly.io work. Unfortunately I don't remember all the details, but effectively fly.io and NodeBB were working under different assumptions as far as local resources were concerned. Upon idle my fly machine would lose track of the fly database, and upon restart it would attempt to restart NodeBB from scratch. Until the idle everything worked fine. But at the end of the day my issue was automatically serving the precompiled Javascript resources that NodeBB presupposes.

Outsourcing the database to something like S3 might fix the problem. Also, fly.io changes very quickly, and I know others were having the same problems with other frameworks, so they may now have a workaround.