k3s/examples/glusterfs/README.md

47 lines
1.6 KiB
Markdown
Raw Normal View History

## Glusterfs
[Glusterfs](http://www.gluster.org) is an open source scale-out filesystem. These examples provide information about how to allow containers use Glusterfs volumes.
The example assumes that the Glusterfs client package is installed on all nodes.
### Prerequisites
Install Glusterfs client package on the Kubernetes hosts.
### Create a POD
The following *volume* spec illustrates a sample configuration.
```js
{
"name": "glusterfsvol",
"glusterfs": {
"endpoints": "glusterfs-cluster",
"path": "kube_vol",
"readOnly": true
}
}
```
The parameters are explained as the followings.
- **endpoints** is endpoints name that represents a Gluster cluster configuration. *kubelet* is optimized to avoid mount storm, it will randomly pick one from the endpoints to mount. If this host is unresponsive, the next Gluster host in the endpoints is automatically selected.
- **path** is the Glusterfs volume name.
- **readOnly** is the boolean that sets the mountpoint readOnly or readWrite.
Detailed POD and Gluster cluster endpoints examples can be found at [v1beta3/](v1beta3/) and [endpoints/](endpoints/)
```shell
# create gluster cluster endpoints
$ kubectl create -f examples/glusterfs/endpoints/glusterfs-endpoints.json
# create a container using gluster volume
$ kubectl create -f examples/glusterfs/v1beta3/glusterfs.json
```
Once that's up you can list the pods and endpoint in the cluster, to verify that the master is running:
```shell
$ kubectl get endpoints
$ kubectl get pods
```
If you ssh to that machine, you can run `docker ps` to see the actual pod and `mount` to see if the Glusterfs volume is mounted.