Note from your TA

[PatNei]

Hi everyone!

Seems like i was wrong about dnssr.
https://docs.docker.com/engine/swarm/networking/#configure-service-discovery
traefik/traefik#913

I found this https://distr.sh/blog/docker-swarm-load-balancing-routing/ and it seems to be working now. specific traefik file. Which would be a problem if you didn’t have traefik.http.services.mygrafana.loadbalancer.swarm=true.

It might be a case of configuration drift on your server. When you get the one click deployment working, see if will magically work by taking the stuff down and up again.

Best of luck!

• P

[AitingLee]

Hi thank you!

We've magically solved the issue by removing dnssr and add these parameters when the nodes join the swarm cluster as the commit shown below. Not sure how exactly it works but we tried one click deployment and update on our production, both works. (For production I need to manually ssh to the server, leave the cluster and rejoin with the same command)

Thanks for the article, it seems like a good reading. I'll try to read and understand our networking better from there :D