Nginx Restart Required

9 months ago

Whenever I redeploy my frontend or backend, all requests to that redeployed service hang until Nginx is restarted. What could cause this behavior? Doesn't happen on my locally hosted VPS with the same containers.

Solved

9 Replies

samgordon

PROOP

9 months ago

322ea2ee-d764-421d-9d11-77dde0980f62

samgordon

PROOP

9 months ago

https://github.com/Loudbooks/PasteBook/blob/master/nginx/default.conf.template

My config is here. The variables are replaced to their internal Railway counterparts, for example backend.railway.internal:8080

brody

EMPLOYEE

9 months ago

background: services on Railway all have dynamic ips that change every deployment.

problem: nginx caches the ip on boot and doesn't reevaluate it, when you deploy the upstream nginx is left making requests to an ip that doesnt go anywhere.

solution: use caddy -

samgordon

PROOP

9 months ago

Ah. Interesting. Thanks!

brody

EMPLOYEE

9 months ago

no problem!

samgordon

PROOP

9 months ago