cappctl

WIP

This tool intends to help solve the chicken and egg problem, Cluster-API from the Kubernetes project uses a Kubernetes cluster to deploy and manage further Kubernetes clusters.. However where does the first cluster come from. cappctl will use plunder to deploy your first cluster, at which point cluster-api will be used to deploy further clusters.

NOTE This tooling has been seperated from the cluster-api-provider-plunder repository do simplify the issues with one project using go modules and the other not, they may be merged in the future.

At this time the initial cappctl tooling can only be used to handle the initial deployment of the first Kubernetes node.

Getting the cluster-api-provider-plunder CLI tooling

Geting the cluster api provider:

go get github.com/plunder-app/cappctl

Building the initial boot strap tooling:

make install

This will install the cappctl tool in the ./bin directory.

Usage

The plunder client configuration will be needed in the current working directory, in order for the tooling to successfully communicate.

Find a server

pldrctl get unleased will return a list of servers that are currently unprovisioned

Bootstrap first node

Pass the Mac address and an IP address to the cappctl tool to bootstrap the first Kubernetes node.

cappctl init-mgmt-cluster -a 192.168.1.2 -m 00:11:22:33:44:55

Destroy a machine

Note If the machine still exists in the plunder deployments, then it will have it's OS reprovisioned.

cappctl destroy -a <machine address>