Cansin Acarer

Jump into the terminal of a CapRover app named testapp. Use Ctrl + P, then Ctrl + Q to exit that terminal without killing the container.

docker exec -it $(docker ps --filter name=srv-captain--testapp -q) /bin/sh

Logs of the deployed application (–follow keeps retrieving the latest events):

docker service logs srv-captain--my-app --since 60m --follow

CapRover logs:

docker service logs captain-captain --since 60m --follow

Logs for Docker itself (if your application keeps getting respawned in new containers, the answer may be in here):

systemctl status -l docker

See nginx records:

docker service ps captain-nginx --no-trunc

Restart CapRover:

docker service update captain-captain --force

docker service scale captain-captain=0
docker service scale captain-captain=1

Restart Docker

sudo systemctl daemon-reload
sudo systemctl restart docker

If your app is running in local but the container is repeatedly crashing on the server, it may be because of a missing dependency only used by the server. For example, if you are testing in local with flask run while the server uses gunicorn, your requirements.txt may be missing some dependencies, as my requirements.txt was missing the following once. I ended up installing these in the local virtual environment as a workaround, just so they are included in each new copy of the requirements.txt after a new pip freeze.

gevent==22.10.1
greenlet==1.1.3.post0
gunicorn==20.1.0

If you are testing locally on a Windows machine, try running the Dockerized version on WSL. The logs on the WSL terminal can contain more information.

docker build -t triggers . && docker run -it triggers

Edit the CapRover nginx config to limit dashboard access to whitelisted IPs:

docker exec -it $(docker ps --filter name=captain-nginx -q) /bin/sh
apt update
apt install nano -y
nano /etc/nginx/conf.d/captain-root.conf

# Captain dashboard at captain.captainroot.domain.com
# ...
    server {
# ...
        # Add this under the line of server_name  captain.***.com;
        # Allow only 169.55.161.198 and deny all others
        allow 169.55.161.198;
        deny all;

If you are getting errors saying you are running out of space, even after you see plenty of free space after running df -h, you might actually be running out of indexes instead. You can see this by running df -i. CapRover cleanup tool doesn’t do a good job removing everything, so instead you can run the following to open up space.

docker system prune -a   # Removes unused containers, networks, images