Caprover: Capatain Duck Duck High availability

[Briston-KS]

Hi there,

Is there any methods by which we can make the CDD server with high availability. For example consider a case in which CDD is deployed in one server and we have applications installed on this server. In worst case scenario consider if the server is rebooted or encounter any issues, the entire application will have a downtime.

I'm checking if there are any methods such as provisioning and clustering the entire CDD servers in high availability such that if one server goes down, the application will be served from the another server without changing the Domain name.

[githubsaturn]

Captain has native support for redundancy. You can create a cluster of webservers and docker swarm automatically puts application on appropriate server (considering load and etc). If a worker server goes down, all apps running on that machine automatically get load balanced on other machines.

Having said this, it doesn't apply to your main node. Main node, i.e., the leader, has meta data about the cluster, has SSL certificates and etc. This is the server where your DNS record will point to. If this server goes down, the cluster is dead.

[Briston-KS]

Thanks for the information.

Is there any option or methods to make the leader node in high availability like using it under Amazon Load balancer or Haproxy etc

[githubsaturn]

I looked into a few options, but couldn't find a "universal" solution that works. I have it on my list to revisit this in the future.

[dottodot]

So does that mean you can't just have multiple managers to make to have high availability or can there only be one leader?

[githubsaturn]

Multiple managers are possible, but there is no point to it. As there will be only one leader anyways.