I am running nodebb on nginx with letsencrypt. I don't see any difference from running it directly.

Hello,

I am trying to setup nodebb with haproxy instead of nginx as frontend. Everything seam to be working however I am getting lots of 403 on /socket.io/ requests.

frontend http-in
mode http
bind 0.0.0.0:80
redirect scheme https code 301 if !{ ssl_fc }

frontend https-in
bind 0.0.0.0:443 ssl crt /etc/letsencrypt/live/test/test.pem
http-response set-header strict-transport-security "max-age=31536000; includeSubDomains"
http-response set-header Content-Security-Policy "default-src 'self' wss: https: *.startech-rd.tk/*; script-src 'self' 'unsafe-inline' 'unsafe-eval' https://code.jquery.com; style-src 'self' 'unsafe-inline' https:; img-src 'self' https://casper.ghost.org/ https://www.gravatar.com/ data:; font-src 'self' https:"
http-response set-header X-XSS-Protection "1; mode=block"
http-response set-header X-Content-Type-Options "nosniff"
http-response set-header Referrer-Policy "no-referrer"
reqadd X-Forwarded-Proto:\ https
acl is_websocket hdr(Upgrade) -i WebSocket
acl is_websocket path_sub -i /socket.io/
use_backend bk_ws if is_websocket
acl acl_comments path_beg -i /comments
use_backend comments if acl_comments

backend comments
mode         http
balance      leastconn
timeout      connect 1s
timeout      server  600s
timeout      queue   600s
option redispatch
retries 3
acl is_woff capture.req.uri -m sub .woff
acl is_ttf capture.req.uri -m sub .ttf
acl is_eot capture.req.uri -m sub .eot
http-response set-header Cache-Control public if is_eot or is_woff or is_ttf
http-response set-header Expires -1 if is_eot or is_woff or is_ttf
http-response set-header Pragma cache if is_eot or is_woff or is_ttf
cookie nodebb insert indirect nocache secure

server node1 10.160.125.81:4567 cookie nodebb_node1 check inter 1000 fastinter 500 rise 2 fall 1
server node2 10.160.125.82:4567 cookie nodebb_node2 check inter 1000 fastinter 500 rise 2 fall 1

backend bk_ws
option redispatch
balance roundrobin
option forwardfor
option httpclose
server node1 10.160.125.81:4567 maxconn 30000 weight 10 cookie ws_node1 check
server node2 10.160.125.82:4567 maxconn 30000 weight 10 cookie ws_node2 check```

I have tried to connect directly without haproxy and the websockets are connecting correctly. However I've seen that using the haproxy the websocket protocol changed from wss to https.

Any suggestions on how to fix this?

@Hillcow Sites-available and sites-enabed does only exist if you install the nginx from the OS repository if you install from any other repository you will only have conf.d which is a good place to put your configuration with the small change that the files in there need to end with .conf extensions.

@absurdsmash

First of all change the server_name for forum.mydomain.com to forum.mydomain.com and restart the nginx. Second of all please show the configuration of default and 22222 so that we can find out which is the proper config for mydomain.com.

@absurdsmash By virtual host I mean your nginx configuration for a different domain/subdomain.

Basically you have an A record for mydomain.com but you missing the A record for forum.mydomain.com. Once that is done and is resolving correctly you will need to add another config file to /etc/nginx/conf.d or /etc/nginx/sites-enabled for forum.mydomain.com

@absurdsmash said in install on wp site:

gital ocean networking s

First of all you don't have any records for forum.mydomain.com that's exactly the DNS error you see in your browser. For using this setting you will have to create a complete new virtualhost configuration.

Only adding as mydomain.com/forum you would need to add the configuration to the mydomain.com virtualhost settings.

@teh_g It works every time in here. Finally no issues.

Does not seam to be lagging for me. If I write codes inside the composer the preview it's slower a little but but I guess cause it needs to parse the code and apply different colors depending on what's inside it. But hey I can wait 2ms until it shows up. It's not like it's lagging tens of secconds or even minutes.

upgraded,

Logs shows some minifier file not found but blog works OK so far.

 [minifier] file not found, node_modules/nodebb-plugin-composer-default/node_modules/screenfull/dist/screenfull.js

here is the error I've told about.

Can you try starting it in developer mode with:

 ./nodebb dev

and give us the output of the startup.