rancher/local-path-provisioner

Local Path Provisioner

Overview

Local Path Provisioner provides a way for the Kubernetes users to utilize the local storage in each node. Based on the user configuration, the Local Path Provisioner will create either hostPath or local based persistent volume on the node automatically. It utilizes the features introduced by Kubernetes Local Persistent Volume feature, but makes it a simpler solution than the built-in local volume feature in Kubernetes.

Compare to built-in Local Persistent Volume feature in Kubernetes

Pros

Dynamic provisioning the volume using hostPath or local.

• Currently the Kubernetes Local Volume provisioner cannot do dynamic provisioning for the local volumes.
• Local based persistent volumes are an experimental feature (example usage).

Cons

1. No support for the volume capacity limit currently.
   1. The capacity limit will be ignored for now.

Requirement

Kubernetes v1.12+.

Deployment

Installation

In this setup, the directory /opt/local-path-provisioner will be used across all the nodes as the path for provisioning (a.k.a, store the persistent volume data). The provisioner will be installed in local-path-storage namespace by default.

• Stable

kubectl apply -f https://raw.githubusercontent.com/rancher/local-path-provisioner/v0.0.36/deploy/local-path-storage.yaml

• Development

kubectl apply -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/deploy/local-path-storage.yaml

Or, use kustomize to deploy.

• Stable

kustomize build "github.com/rancher/local-path-provisioner/deploy?ref=v0.0.36" | kubectl apply -f -

• Development

kustomize build "github.com/rancher/local-path-provisioner/deploy?ref=master" | kubectl apply -f -

After installation, you should see something like the following:

$ kubectl -n local-path-storage get pod
NAME                                     READY     STATUS    RESTARTS   AGE
local-path-provisioner-d744ccf98-xfcbk   1/1       Running   0          7m

Check and follow the provisioner log using:

kubectl -n local-path-storage logs -f -l app=local-path-provisioner

Usage

Create a hostPath backend Persistent Volume and a pod uses it:

kubectl create -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/examples/pvc/pvc.yaml
kubectl create -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/examples/pod/pod.yaml

Or, use kustomize to deploy them.

kustomize build "github.com/rancher/local-path-provisioner/examples/pod?ref=master" | kubectl apply -f -

You should see the PV has been created:

$ kubectl get pv
NAME                                       CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS    CLAIM                    STORAGECLASS   REASON    AGE
pvc-bc3117d9-c6d3-11e8-b36d-7a42907dda78   2Gi        RWO            Delete           Bound     default/local-path-pvc   local-path               4s

The PVC has been bound:

$ kubectl get pvc
NAME             STATUS    VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS   AGE
local-path-pvc   Bound     pvc-bc3117d9-c6d3-11e8-b36d-7a42907dda78   2Gi        RWO            local-path     16s

And the Pod started running:

$ kubectl get pod
NAME          READY     STATUS    RESTARTS   AGE
volume-test   1/1       Running   0          3s

Write something into the pod

kubectl exec volume-test -- sh -c "echo local-path-test > /data/test"

Now delete the pod using

kubectl delete -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/examples/pod/pod.yaml

After confirm that the pod is gone, recreated the pod using

kubectl create -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/examples/pod/pod.yaml

Check the volume content:

$ kubectl exec volume-test -- sh -c "cat /data/test"
local-path-test

Delete the pod and pvc

kubectl delete -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/examples/pod/pod.yaml
kubectl delete -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/examples/pvc/pvc.yaml

Or, use kustomize to delete them.

kustomize build "github.com/rancher/local-path-provisioner/examples/pod?ref=master" | kubectl delete -f -

The volume content stored on the node will be automatically cleaned up. You can check the log of local-path-provisioner-xxx for details.

Now you've verified that the provisioner works as expected.

Configuration

Customize the ConfigMap

The configuration of the provisioner is a json file config.json, a Pod template helperPod.yaml and two bash scripts setup and teardown, stored in a config map, e.g.:

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":["/opt/local-path-provisioner"]
                },
                {
                        "node":"yasker-lp-dev1",
                        "paths":["/opt/local-path-provisioner", "/data1"]
                },
                {
                        "node":"yasker-lp-dev3",
                        "paths":[]
                }
                ]
        }
  setup: |-
        #!/bin/sh
        set -eu
        mkdir -m 0777 -p "$VOL_DIR"
  teardown: |-
        #!/bin/sh
        set -eu
        rm -rf "$VOL_DIR"
  helperPod.yaml: |-
        apiVersion: v1
        kind: Pod
        metadata:
          name: helper-pod
        spec:
          priorityClassName: system-node-critical
          tolerations:
            - key: node.kubernetes.io/disk-pressure
              operator: Exists
              effect: NoSchedule
          containers:
          - name: helper-pod
            image: busybox

The helperPod is allowed to run on nodes experiencing disk pressure conditions, despite the potential resource constraints. When it runs on such a node, it can carry out specific cleanup tasks, freeing up space in PVCs, and resolving the disk-pressure issue.

config.json

Definition

nodePathMap is the place user can customize where to store the data on each node.

1. If one node is not listed on the nodePathMap, and Kubernetes wants to create volume on it, the paths specified in DEFAULT_PATH_FOR_NON_LISTED_NODES will be used for provisioning.
2. If one node is listed on the nodePathMap, the specified paths in paths will be used for provisioning.
   1. If one node is listed but with paths set to [], the provisioner will refuse to provision on this node.
   2. If more than one path was specified, the path would be chosen randomly when provisioning.

sharedFileSystemPath allows the provisioner to use a filesystem that is mounted on all nodes at the same time. In this case all access modes are supported: ReadWriteOnce, ReadOnlyMany and ReadWriteMany for storage claims.

storageClassConfigs is a map from storage class names to objects containing nodePathMap or sharedFilesystemPath, as described above.

In addition volumeBindingMode: Immediate can be used in StorageClass definition.

Please note that nodePathMap, sharedFileSystemPath, and storageClassConfigs are mutually exclusive. If sharedFileSystemPath or storageClassConfigs are used, then nodePathMap must be set to [].

The setupCommand and teardownCommand allow you to specify the path to binary files in helperPod that will be called when creating or deleting pvc respectively. This can be useful if you need to use distroless images for security reasons. See the examples/distroless directory for an example. A binary file can take the following parameters:

The setupCommand and teardownCommand have higher priority than the setup and teardown scripts from the ConfigMap.

Rules

The configuration must obey following rules:

1. config.json must be a valid json file.
2. A path must start with /, a.k.a an absolute path.
3. Root directory(/) is prohibited.
4. No duplicate paths allowed for one node.
5. No duplicate node allowed.

Scripts setup and teardown and the helperPod.yaml template

• The setup script is run before the volume is created, to prepare the volume directory on the node.
• The teardown script is run after the volume is deleted, to cleanup the volume directory on the node.
• The helperPod.yaml template is used to create a helper Pod that runs the setup or teardown script.

By default, the helper pod template is validated before it is accepted. The template may customize the helper container image and regular pod settings such as priorityClassName and tolerations, but it must not define its own securityContext, volumes, volumeMounts, initContainers, ephemeralContainers, envFrom, env.valueFrom, lifecycle, livenessProbe, readinessProbe, startupProbe, host namespaces, spec.nodeName, or spec.serviceAccountName.

If you intentionally need a privileged helper pod template, you must opt in explicitly by setting --allow-unsafe-helper-pod-template or ALLOW_UNSAFE_HELPER_POD_TEMPLATE=true on the provisioner deployment. This disables those safety checks and should only be used for trusted, administrator-managed configurations.

The scripts receive their input as environment variables:

Reloading

The provisioner supports automatic configuration reloading. Users can change the configuration using kubectl apply or kubectl edit with config map local-path-config. There is a delay between when the user updates the config map and the provisioner picking it up. In order for this to occur for updates made to the helper pod manifest, the following environment variable must be added to the provisioner container. If not, then the manifest used for the helper pod will be the same as what was in the config map when the provisioner was last restarted/deployed.

- name: CONFIG_MOUNT_PATH
  value: /etc/config/

When the provisioner detects the configuration changes, it will try to load the new configuration. Users can observe it in the log

time="2018-10-03T05:56:13Z" level=debug msg="Applied config: {"nodePathMap":[{"node":"DEFAULT_PATH_FOR_NON_LISTED_NODES","paths":["/opt/local-path-provisioner"]},{"node":"yasker-lp-dev1","paths":["/opt","/data1"]},{"node":"yasker-lp-dev3"}]}"

If the reload fails, the provisioner will log the error and continue using the last valid configuration for provisioning in the meantime.

time="2018-10-03T05:19:25Z" level=error msg="failed to load the new config file: fail to load config file /etc/config/config.json: invalid character '#' looking for beginning of object key string"

time="2018-10-03T05:20:10Z" level=error msg="failed to load the new config file: config canonicalization failed: path must start with / for path opt on node yasker-lp-dev1"

time="2018-10-03T05:23:35Z" level=error msg="failed to load the new config file: config canonicalization failed: duplicate path /data1 on node yasker-lp-dev1

time="2018-10-03T06:39:28Z" level=error msg="failed to load the new config file: config canonicalization failed: duplicate node yasker-lp-dev3"

The same behavior applies to helperPod.yaml: if a reloaded template violates the helper pod safety checks, the provisioner logs the validation error and keeps using the last valid helper pod template.

Volume Types

To specify the type of volume you want the provisioner to create, add either of the following annotations;

• PVC:

annotations:
  volumeType: <local or hostPath>

• StorageClass:

annotations:
  defaultVolumeType: <local or hostPath>

A few things to note; the annotation for the StorageClass will apply to all volumes using it and is superseded by the annotation on the PVC if one is provided. If neither of the annotations was provided then we default to hostPath.

Storage classes

If more than one paths are specified in the nodePathMap the path is chosen randomly. To make the provisioner choose a specific path, use a storageClass defined with a parameter called nodePath. Note that this path should be defined in the nodePathMap.

By default the volume subdirectory is named using the template {{ .PVName }}_{{ .PVC.Namespace }}_{{ .PVC.Name }} which make the directory specific to the PV instance. The template can be changed using the pathPattern parameter which is interpreted as a go template. The template has access to the PV name using the PVName variable and the PVC metadata object, including labels and annotations, with the PVC variable.

When pathPattern is set, the rendered path must start with {{ .PVC.Namespace }}/{{ .PVC.Name }}/ and must not contain directory traversal (for example ../).

If you need to keep an existing pathPattern that does not follow the prefix requirement, you can opt out by setting allowUnsafePathPattern: "true" on the StorageClass (either in parameters or metadata.annotations). When enabled, the provisioner will skip these validations.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: ssd-local-path
provisioner: rancher.io/local-path
parameters:
  nodePath: /data/ssd
  pathPattern: "{{ .PVC.Namespace }}/{{ .PVC.Name }}/"
volumeBindingMode: WaitForFirstConsumer
reclaimPolicy: Delete

Here the provisioner will use the path /data/ssd with a subdirectory per namespace and PVC when storage class ssd-local-path is used.

Node Affinity Key

By default, PersistentVolumes created by the provisioner use the kubernetes.io/hostname label for node affinity. This ensures the volume is only accessible on the node where it was provisioned.

In environments where node hostnames are not stable (e.g., VMs managed by orchestrators that assign new hostnames on recreation), the PV node affinity can reference a stale hostname, making the volume unschedulable. The nodeAffinityKey StorageClass parameter allows you to specify a different node label key that remains stable across node restarts or recreation.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: local-path-stable
provisioner: rancher.io/local-path
parameters:
  nodeAffinityKey: my.domain/stable-node-id
volumeBindingMode: WaitForFirstConsumer
reclaimPolicy: Delete

When using this parameter, make sure every node in the cluster has the specified label set. The provisioner reads the label value from the selected node at provisioning time and writes it into the PV's nodeAffinity field.

If the parameter is not specified, the default kubernetes.io/hostname behavior is preserved.

Upgrade note: Existing PVs created before adding nodeAffinityKey retain their original kubernetes.io/hostname affinity. The provisioner handles deletion of both old and new PVs transparently. Only newly provisioned PVs will use the custom key.

Uninstall

Before uninstallation, make sure the PVs created by the provisioner have already been deleted. Use kubectl get pv and make sure no PV with StorageClass local-path.

To uninstall, execute:

• Stable

kubectl delete -f https://raw.githubusercontent.com/rancher/local-path-provisioner/v0.0.36/deploy/local-path-storage.yaml

• Development

kubectl delete -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/deploy/local-path-storage.yaml

Debug

it providers a out-of-cluster debug env for developers

debug

git clone https://github.com/rancher/local-path-provisioner.git
cd local-path-provisioner
go build
kubectl apply -f debug/config.yaml
./local-path-provisioner --debug start --service-account-name=default

example

Usage

clear

kubectl delete -f debug/config.yaml

License

Copyright (c) 2014-2020 Rancher Labs, Inc.

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.