30f856b417
* Serve assets from /assets, remove /images endpoint * Migrate images to assets so we have a place for general assets
Examples
Examples contains bootcfg data directories showcasing different network-bootable baremetal clusters. These examples work with libvirt VMs setup with known hardware attributes by scripts/libvirt.
Clusters
| Name | Description | Type | Docs |
|---|---|---|---|
| etcd-small | Cluster with 1 etcd node, 4 proxies | cloud config | reference |
| etcd-large | Cluster with 3 etcd nodes, 2 proxies | cloud config | reference |
Usage
First, read about the bootcfg config service. Then setup the config service and network boot environment using the libvirt guide.
Create 5 libvirt VM nodes on the docker0 bridge.
./scripts/libvirt create # create node1 ... node5
Generally, you just need to run the coreos/bootcfg container and the coreos/dnsmasq container (which runs network boot services).
For coreos/bootcfg, mount one of the example directories with -v $PWD/examples/some-example:/data:Z and provide the -data-path if the mount location differs from /data.
Reboot to see nodes boot with PXE and create a cluster.
./scripts/libvirt reboot
Clean up by powering off and destroying the VMs.
./scripts/libvirt shutdown # graceful
./scripts/libvirt poweroff # non-graceful
./scripts/libvirt destroy