DigitalOcean droplet: 503 Service Temporarily Unavailable

Newbie here – trying this out. On DigitalOcean deployed the droplet, then through Droplet Console (or ssh) can login via root and initially I see the request for a domain name – so I’ve provided that. After that the IP address for the droplet can be pinged, so it’s alive. But when I click on the blue button on the page saying “Get Started”/“Getting started after deploying Gramps Web”, it’s returning a " 503 Service Temporarily Unavailable". I’ve followed steps 1 and 2 on DigitalOcean - Gramps Web, maybe step 3 is going wrong.

Could you please give me a few more details on Step 3? I’ve pointed the nameservers on my domain to the digital ocean nameservers. What else do I need to do there?

Wouldn’t your domain nameservers point at the IP rather than the DigitalOcean nameservers?

DigitalOceans is just providing a fixed IP address, not DNS service.

Will try as well and let you know. Can’t exclude some recent updates of GW need modification of the DO app.

1 Like

well, for the moment just going to the IP address directly by clicking on the “Getting Started” button isn’t working. Before I go to gramps.mydomainname.com I guess we should see if just going to the IP address is working, but maybe I am missing something…not sure.

I wouldn’t be sure about that, since it’s using acme-companion based on nginx. 503 is the typical status when nginx is unhappy.

Please try with a domain and let me know. You can use DuckDNS; it’s also possible to change it later.

Had to look that up!

acme-companion is a lightweight companion container for nginx-proxy

It handles the automated creation, renewal and use of SSL certificates for proxied Docker containers through the ACME protocol.

So you’re saying the SSL certificate has to validated against the domain name?

Which means the gramps. mydomainname .com subdomain has to both be pointed at the IP address and that DNS entry has to have propagated through to whatever nameserver the nginx-proxy is using (which might take 48 hours) before it will be happy?

The problem is that this setup uses nginx with reverse proxy to Gramps Web, and it uses environment variables with the fullly qualified domain name to do the routing, which are set by the first login script, as described in the documentation.

If you can ping the server but get 503, it means the reverse proxy doesn’t work, which is not surprising because the nginx rewrite rule doesn’t catch the right host when you use the IP address.

Yes, I can ping the IP address, and it does say “nginx” below the " 503 Service Temporarily Unavailable", so you are probably correct…

So now I updated my GoDaddy A record so that gramps.mydomain.com points to the IP address and so now if I go to gramps.mydomain.com I see the same message “503 Service Temporarily Unavailable”. Maybe it needs some more time, but what else could I do for debugging?

Thanks for the suggestions

1 Like

Update – seemed to work! It seems I had a mistake when the startup script said enter the domain name you will use for the Gramps Web: I had put in mydomain.com instead of gramps.mydomain.com

I changed that – went through the startup script again…and now at least I can see the UI on the web…

thanks a lot for all your suggestions

2 Likes

Great! Please let us know if you encounter any issues during setup. As said, it could be that some recent updates require us to update the DO app as well as I haven’t tested it in a while.

1 Like