Deploying resources to multiple servers with HCS

Multiple Servers

A guide on how to deploy a resource to multiple servers with HCS.

With this feature, You could deploy the same application to multiple servers, add a load balancer in front of them and you will get a highly available application.

Requirements

• Each server should be added to HCS, validated and reachable.

• Each server (and the optional build server) should be the same architecture (AMD64, ARM).

• You must push the built image to a Docker Registry. HCS automates the process, you just need to login to the registry on the server.

How to use?

When you configure (or already configured) an application, you selected a server / network where it deploys. This will be your main server.

Any additional servers must be set in the Servers menu, simply by clicking on it.

Now every time you redeploy, restart or stop the application, the action will be done on all servers.

If the deploy needs a build process, it will be executed on the main server (or on the build server if you have on). The deploy process will upload the built image to the Docker Registry and only after all other servers will be notified to pull and deploy this image.

How to configure a loadbalancer?

At the moment, it is not automated. So you have to manually setup a loadbalancer. There are two ways to use.

Port mapping to host

If you set Ports Mappings for your application, so one port from the docker container will mapped to a port on the host server, all you need to do is to:

1. Set all the IP:PORT as a destination in your loadbalancer.

2. Remove any domains from the Domains field in HCS.

In this case, HCS Proxy is not used as you can reach the application with IP:PORT

Using a domain

In this case, you need to set the loadbalancer domain with HTTP, (not HTTPS) in the Domains field, and then set the proper configuration for your loadbalancer, with SSL termination.

With this configuration, you can use several domains with one loadbalancer.