Rk8s - A Kubernetes Distribution with Rust

For now, we are woking on a deploy tool helps you set up a working kubernetes cluster in just a few steps.

Usage

Prerequisite:

CentOS Stream 9: the currently tested working distro, since the rust container runtime youki has a limit on linux kernel version.

gcc: compiler which rustc depends on.

cargo command available: which means you need to have rust installed.

cfssl: If you do not have one installed, you can run rk8s install cfssl after the configuration is generated.

First, clone & build the crate:

git clone https://github.com/open-rust-initiative/rk8s.git
cd rk8s
cargo build
cp target/debug/rk8s /usr/bin

Then you should now have a working rk8s ready for deploying.

The rk8s command needs root privilege.

Deploying outside the cluster:

1. rk8s generate config will generate a folder named rk8s under /root directory.

2. Change the content in /root/rk8s/cfg/config.yaml, specify the machines' IP addresses and their according roles (master or worker) in instance_hosts filed, if your deploying machine (the machine running rk8s) will be outside of cluster, then instance_ip and instance_name fields are irrelevant.

3. ssh-keygen to generate a key for ssh connection across machines, and ssh-copy-id -i <path/to/.pub> root@<IP address> notifies machines to be deployed.

4. rk8s install cfssl to install cfssl-related tools for later use.

5. Run rk8s deploy.

Then you should have a working cluster, ssh to the master node and run kubectl get nodes, you should see the master node is ready.

Deploying inside the cluster:

1. Run rk8s generate config on every machine to be deployed.

2. Run rk8s precheck to turn off selinux, firewall and so forth.

3. Change the content in /root/rk8s/cfg/config.yaml, specify the machines' IP addresses and their according roles (master or worker) in instance_hosts filed, instance_ip and instance_name fields need to be set correctly according to the IP and hostname of the current machine.

4. Run rk8s install etcd on each instance.

5. For masters run:

   $ rk8s install kubernetes
   $ rk8s install docker

6. Run rk8s init on master node.

7. Run rk8s join on worker node.

After rk8s finishes its job, the cluster is now set, the deploying machine is inside the cluster.