Usernetes (Gen2) deploys a Kubernetes cluster inside Rootless Docker, so as to mitigate potential container-breakout vulnerabilities.
Note
Usernetes (Gen2) has significantly diverged from the original Usernetes (Gen1), which did not require Rootless Docker to be installed on hosts.
See the
gen1
branch for the original Usernetes (Gen1).
Usernetes (Gen2) is similar to Rootless kind
and Rootless minikube,
but Usernetes (Gen 2) supports creating a cluster with multiple hosts.
- Cluster configuration: kubeadm
- CRI: containerd
- OCI: runc
- CNI: Flannel
- One of the following host operating system:
Host operating system | Minimum version |
---|---|
Ubuntu (recommended) | 22.04 |
Rocky Linux | 9 |
AlmaLinux | 9 |
Fedora | (?) |
- One of the following container engines:
Container Engine | Minimum version |
---|---|
Rootless Docker (recommended) | v20.10 |
Rootless Podman | v4.x |
Rootless nerdctl | v1.6 |
curl -o install.sh -fsSL https://get.docker.com
sudo sh install.sh
dockerd-rootless-setuptool.sh install
- systemd lingering:
sudo loginctl enable-linger $(whoami)
- cgroup v2 delegation:
sudo mkdir -p /etc/systemd/system/[email protected]
sudo tee /etc/systemd/system/[email protected]/delegate.conf <<EOF >/dev/null
[Service]
Delegate=cpu cpuset io memory pids
EOF
sudo systemctl daemon-reload
- Kernel modules:
sudo tee /etc/modules-load.d/usernetes.conf <<EOF >/dev/null
br_netfilter
vxlan
EOF
sudo systemctl restart systemd-modules-load.service
- sysctl:
sudo tee /etc/sysctl.d/99-usernetes.conf <<EOF >/dev/null
net.ipv4.conf.default.rp_filter = 2
EOF
sudo sysctl --system
Use scripts in ./init-host
for automating these steps.
See make help
.
# Bootstrap a cluster
make up
make kubeadm-init
make install-flannel
# Enable kubectl
make kubeconfig
export KUBECONFIG=$(pwd)/kubeconfig
kubectl get pods -A
# Multi-host
make join-command
scp join-command another-host:~/usernetes
ssh another-host make -C ~/usernetes up kubeadm-join
make sync-external-ip
# Debug
make logs
make shell
make kubeadm-reset
make down-v
kubectl taint nodes --all node-role.kubernetes.io/control-plane-
The container engine defaults to Docker.
To change the container engine, set export CONTAINER_ENGINE=podman
or export CONTAINER_ENGINE=nerdctl
.
- Node ports cannot be exposed automatically. Edit
docker-compose.yaml
for exposing additional node ports. - Most of host files are not visible with
hostPath
mounts. Editdocker-compose.yaml
for mounting additional files. - Some volume drivers such as
nfs
do not work.
When CONTAINER_ENGINE
is set to nerdctl
, bypass4netns can be enabled for accelerating connect(2)
syscalls.
The acceleration currently does not apply to VXLAN packets.
containerd-rootless-setuptool.sh install-bypass4netnsd
export CONTAINER_ENGINE=nerdctl
make up