Kubernetes

The Kubernetes integration backs up cluster resources and persistent volumes. It provides two connectors accessible via two URI schemes:

URI scheme	What it backs up
`k8s://`	Kubernetes manifests and resource state across namespaces.
`k8s+csi://`	Persistent volume contents via CSI driver snapshots.

Requirements

Plakar v1.1.0-beta or later.
kubectl proxy running and accessible.
A CSI driver with snapshot support and a configured VolumeSnapshotClass for CSI-based PVC backups.

Typical use cases

Namespace or resource-level restore from manifest snapshots.
Incident investigation by browsing cluster state at a point in time.
Persistent volume backup and cross-environment data portability.

Installation

Note: Installing pre-built packages requires authentication with Plakar. See Login to Plakar to unlock features.

Install the Kubernetes package:

$ plakar pkg add k8s

Verify the installation:

$ plakar pkg list

Prerequisites

A working Go toolchain compatible with your version of Plakar.

Build and install the Kubernetes package:

$ plakar pkg build k8s
$ plakar pkg add ./k8s_v1.1.0-beta_linux_amd64.ptar

Verify the installation:

$ plakar pkg list

To check installed packages:

$ plakar pkg list

To upgrade, remove and reinstall:

$ plakar pkg rm k8s && plakar pkg add k8s

Existing configurations are preserved during upgrade.

Manifest backup and restore

The k8s:// connector fetches all Kubernetes resources across the cluster and stores them as a Plakar snapshot. This enables browsing, diffing, and restoring cluster configuration at any level of granularity — full cluster, single namespace, or individual resource.

Snapshots include resource status metadata, making them useful for incident investigation — you can browse the Plakar UI to inspect the state of deployments, nodes, and other resources at any point in time.

flowchart LR subgraph Source[Kubernetes Cluster] api@{ shape: rect, label: "API Server" } end subgraph Plakar[Plakar] Connector@{ shape: rect, label: "Fetch manifests via
kubectl proxy" } Transform@{ shape: rect, label: "Encrypt & deduplicate" } Connector --> Transform end Source --> Connector Store@{ shape: cyl, label: "Kloset Store" } Transform --> Store classDef sourceBox fill:#ffe4e6,stroke:#cad5e2,stroke-width:1px classDef brandBox fill:#524cff,color:#ffffff classDef storeBox fill:#dbeafe,stroke:#cad5e2,stroke-width:1px class Source sourceBox class Plakar brandBox class Store storeBox linkStyle default stroke-dasharray: 9,5,stroke-dashoffset: 900,animation: dash 25s linear infinite;

Start a proxy to the cluster:

$ kubectl proxy
Starting to serve on 127.0.0.1:8001

Back up manifests

Back up all resources across the entire cluster:

$ plakar backup k8s://localhost:8001

Back up resources in a specific namespace:

$ plakar backup k8s://localhost:8001/foo

Restore manifests

Restore all StatefulSet resources in the foo namespace:

$ plakar restore -to k8s://localhost:8001 abcd:/foo/apps/StatefulSet

Persistent volume backup and restore (CSI)

The k8s+csi:// connector backs up the contents of persistent volumes by creating a VolumeSnapshot, mounting it in a temporary pod running a helper importer, and ingesting the data into a Kloset store. The snapshot is deleted from the cluster once ingestion completes.

Restore works in reverse: data is written into a target PVC using the same helper pod mechanism. The target can be an existing PVC or a freshly created one.

flowchart LR subgraph Source[Kubernetes Cluster] pvc@{ shape: cyl, label: "PVC" } snap@{ shape: rect, label: "VolumeSnapshot" } pvc --> snap end subgraph Plakar[Plakar] Connector@{ shape: rect, label: "Ingest via
helper pod" } Transform@{ shape: rect, label: "Encrypt & deduplicate" } Connector --> Transform end snap --> Connector Store@{ shape: cyl, label: "Kloset Store" } Transform --> Store classDef sourceBox fill:#ffe4e6,stroke:#cad5e2,stroke-width:1px classDef brandBox fill:#524cff,color:#ffffff classDef storeBox fill:#dbeafe,stroke:#cad5e2,stroke-width:1px class Source sourceBox class Plakar brandBox class Store storeBox linkStyle default stroke-dasharray: 9,5,stroke-dashoffset: 900,animation: dash 25s linear infinite;

Back up a PVC

$ plakar backup -o volume_snapshot_class=my-snapclass k8s+csi://localhost:8001/storage/my-pvc

Restore a PVC

Restore into a new, empty PVC:

$ kubectl create -f -
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: pristine
  namespace: storage
spec:
  resources:
    requests:
      storage: 1Gi
  accessModes:
    - ReadWriteOnce

$ plakar restore -to k8s+csi://localhost:8001/storage/pristine abcdef:

Restore into an existing PVC by referencing it in the same way. The target PVC must have sufficient capacity.

Options

Option	Required	Description
`volume_snapshot_class`	Yes	Name of the `VolumeSnapshotClass` to use for CSI snapshots.
`kubelet_image`	No	Container image for the helper pod. Defaults to a recent kubelet image.

Limitations and scope

What is captured

All Kubernetes resource manifests and status metadata (k8s://).
Persistent volume contents for CSI-backed PVCs (k8s+csi://).

What is not captured

Non-CSI volumes are not yet supported for PVC backups.
Node-level configuration (OS, kubelet config, network setup).
In-flight workload state (open connections, in-memory data).

Snapshot consistency

Manifest snapshots reflect the state of the API server at the time of backup. For PVCs, consistency depends on the CSI driver and whether the workload was quiesced before the snapshot was taken.

Book a demo