(Caddy) Self-Hosted AWS EC2 Taiga Issues After Start/Stop Server

Troubleshooting Self hosted Taiga
1

Hello! I’ve been enjoying the self-hosted version of Caddy via AWS for the last few days or so…and I am using the Docker version. Ever since I stopped, and started, my EC2 instance, I can no longer get the app to work. Network wise, everything seems fine, and I’ve double checked everything on the AWS side.

I notice, when running docker ps, than the Caddy file is just continuously restarting. I never saw this issue before, and everything else shows Healthy and Running (for docker containers), but this darn Caddy docker container appears to keep restarting.

I’ve recreated the Caddyfile, though it uses the same exact parameters as before, to make sure my domain works.

Is there a way to kill just the Caddy container, and recreate that, instead of everything (I don’t want to lose my DB, stories, users, etc. that have been working in there for days now)? I am not super familiar with Caddy, I just know that when I setup the Caddyfile properly, I was able to access my app just fine from my sub domain.

I’m not sure what is causing the Caddy container to stay stuck restarting, but I assume it has something to do with Stopping / Starting my AWS EC2 instance that Taiga’s been hosted in.

As part of trying to resolve this, I’ve assigned a static (Elastic) IP to my EC2 container as well, which should now never change, but I think the IP did change before that when I did the Stop/Start, so I figure if I assign an Elastic IP, I wouldn’t have to worry about this anymore. However, I need to be able to get this thing back up without restarting the entire Taiga installation process.

Any thoughts? Is it a cert related to Caddy that is recognizing the changed (now twice) IP address and I need to configure this? Where could I?

Thanks a ton, y’all.

2

Hi there!

The official docker-compose from Taiga (taiga-docker/docker-compose.yml at main · taigaio/taiga-docker · GitHub) does not use Caddy, so not a 100% sure where you got that from (unless it’s a modification you made to the docker-compose file).

In any case, I will be assuming that you are using a docker-compose file for the rest of this post.

You can target specific services/containers using docker-compose, for example: docker compose up -d --force-recreate <service_name>. In our official docker-compose file, the service equivalent to Caddy would be nginx, which is called taiga-gateway in the compose file. So, to recreate and restart that container, we would do: docker compose up -d --force-recreate taiga-gateway.

Still, we use docker volumes on our compose files, so unless you specifically delete them or something goes wrong, your data should remain even if you recreate/restart the services.

I hope this helps!

Best regards,