ZFS example

[killfill]

Hi all :wave:

Just adding a tip in here, in case it saves a couple of minutes for people wanted to use ZFS with this local-path-provisioner. At least it will for me, when i forget it :grimacing:

kind: ConfigMap
apiVersion: v1
metadata:
  name: local-path-config
  namespace: local-path-storage
data:
  config.json: |-
        {
                "nodePathMap":[
                {
                        "node":"DEFAULT_PATH_FOR_NON_LISTED_NODES",
                        "paths":["/tank/"]
                }
                ]
        }
  setup: |-
        #!/bin/sh
        while getopts "m:s:p:" opt
        do
            case $opt in
                p)
                absolutePath=$OPTARG
                ;;
                s)
                sizeInBytes=$OPTARG
                ;;
                m)
                volMode=$OPTARG
                ;;
            esac
        done
        #volMode=Filesystem
        zfs create -o quota=${sizeInBytes} ${absolutePath:1}
  teardown: |-
        #!/bin/sh
        while getopts "m:s:p:" opt
        do
            case $opt in
                p)
                absolutePath=$OPTARG
                ;;
                s)
                sizeInBytes=$OPTARG
                ;;
                m)
                volMode=$OPTARG
                ;;
            esac
        done
        zfs destroy ${absolutePath:1}
  helperPod.yaml: |-
        apiVersion: v1
        kind: Pod
        metadata:
          name: helper-pod
        spec:
          containers:
          - name: helper-pod
            image: srueg/zfs-utils
            securityContext:
              privileged: true

[killfill]

Hm.. ok im sorry. That actually did not work.

When creating or destroying ZFS datasets from a container like the podHelper's one, we need to mount the directory as 'Bidirectional', so the changes can propagate to the Host.

This is maybe related to https://github.com/rancher/local-path-provisioner/issues/165

[killfill]

I guess this can be done enabling the mountPropagation=Bidirectional directly, near here, but that would require the pod helper to run privileged, thus probably would need to be enabled via a flag.

Anyway, another aproach could be something like this?

In case anyone cares, just uploaded a docker image with that change.

With the followin config, ZFS does actually work now =)

kind: ConfigMap
apiVersion: v1
metadata:
  name: local-path-config
  namespace: local-path-storage
data:
  config.json: |-
        {
                "nodePathMap":[
                {
                        "node":"DEFAULT_PATH_FOR_NON_LISTED_NODES",
                        "paths":["/tank/"]
                }
                ]
        }
  setup: |-
        #!/bin/sh
        while getopts "m:s:p:" opt
        do
            case $opt in
                p)
                absolutePath=$OPTARG
                ;;
                s)
                sizeInBytes=$OPTARG
                ;;
                m)
                volMode=$OPTARG
                ;;
            esac
        done
        #volMode=Filesystem
        zfs create -o quota=${sizeInBytes} ${absolutePath:1}
        chmod g+w ${absolutePath}
  teardown: |-
        #!/bin/sh
        while getopts "m:s:p:" opt
        do
            case $opt in
                p)
                absolutePath=$OPTARG
                ;;
                s)
                sizeInBytes=$OPTARG
                ;;
                m)
                volMode=$OPTARG
                ;;
            esac
        done
        zfs destroy ${absolutePath:1}
  helperPod.yaml: |-
        apiVersion: v1
        kind: Pod
        metadata:
          name: helper-pod
        spec:
          containers:
          - name: helper-pod
            image: srueg/zfs-utils
            securityContext:
              privileged: true
            volumeMounts:
            - mountPath: /tank
              name: data
              mountPropagation: Bidirectional

[markusressel]

Thx for the effort, looking forward to trying this out in the coming days :nerd_face:

[mgoltzsche]

PR #166 allows to specify the mountPropagation within the helper Pod template.

[haslersn]

What needs to be done in order to integrate this with the volume snapshot API?

[innobead]

I guess this can be done enabling the mountPropagation=Bidirectional directly, near here, but that would require the pod helper to run privileged, thus probably would need to be enabled via a flag.

Anyway, another aproach could be something like this?

In case anyone cares, just uploaded a docker image with that change.

With the followin config, ZFS does actually work now =)

kind: ConfigMap
apiVersion: v1
metadata:
  name: local-path-config
  namespace: local-path-storage
data:
  config.json: |-
        {
                "nodePathMap":[
                {
                        "node":"DEFAULT_PATH_FOR_NON_LISTED_NODES",
                        "paths":["/tank/"]
                }
                ]
        }
  setup: |-
        #!/bin/sh
        while getopts "m:s:p:" opt
        do
            case $opt in
                p)
                absolutePath=$OPTARG
                ;;
                s)
                sizeInBytes=$OPTARG
                ;;
                m)
                volMode=$OPTARG
                ;;
            esac
        done
        #volMode=Filesystem
        zfs create -o quota=${sizeInBytes} ${absolutePath:1}
        chmod g+w ${absolutePath}
  teardown: |-
        #!/bin/sh
        while getopts "m:s:p:" opt
        do
            case $opt in
                p)
                absolutePath=$OPTARG
                ;;
                s)
                sizeInBytes=$OPTARG
                ;;
                m)
                volMode=$OPTARG
                ;;
            esac
        done
        zfs destroy ${absolutePath:1}
  helperPod.yaml: |-
        apiVersion: v1
        kind: Pod
        metadata:
          name: helper-pod
        spec:
          containers:
          - name: helper-pod
            image: srueg/zfs-utils
            securityContext:
              privileged: true
            volumeMounts:
            - mountPath: /tank
              name: data
              mountPropagation: Bidirectional

@killfill I would suggest you can contribute back to the examples that will be great. Thanks.

[killfill]

Hi @innobead,

Thanks for your comment, but i think for this to work it needs the bidirectional mounting thing enabled. May be a good idea to wait for that first. What do you think?

[innobead]

Hi @innobead,

Thanks for your comment, but i think for this to work it needs the bidirectional mounting thing enabled. May be a good idea to wait for that first. What do you think?

Sounds good to me, let's wait for #165 to resolve first.

[derekbit]

@killfill Looks good. Are you ready for submitting a PR for ZFS example?

[killfill]

Hey @derekbit , well I would love to.

But I think some changes needs to be done previously. Im actually using these changes for my little setup, which works, but im not sure its going into the right direction.

[killfill]

All right, adding the example file as a PR. It actually needs the head-master tag, (and probably the next release v0.0.22).

Hope it helps.

[derekbit]

@killfill Sorry for the late reply. The PR Looks good. I will test it before merging it. Thank you!