Skip to content

Latest commit

 

History

History
140 lines (107 loc) · 5.85 KB

ceph-csi-drivers.md

File metadata and controls

140 lines (107 loc) · 5.85 KB
Raw
title weight indent
Ceph CSI
3200
true

Ceph CSI Drivers

There are two CSI drivers integrated with Rook that will enable different scenarios:

  • RBD: This driver is optimized for RWO pod access where only one pod may access the storage
  • CephFS: This driver allows for RWX with one or more pods accessing the same storage

The drivers are enabled automatically with the Rook operator. They will be started in the same namespace as the operator when the first CephCluster CR is created.

For documentation on consuming the storage:

Supported Versions

The supported Ceph CSI version is 3.3.0 or greater with Rook. Refer to ceph csi releases for more information.

Static Provisioning

Both drivers also support the creation of static PV and static PVC from existing RBD image/CephFS volume. Refer to static PVC for more information.

Configure CSI Drivers in non-default namespace

If you've deployed the Rook operator in a namespace other than "rook-ceph", change the prefix in the provisioner to match the namespace you used. For example, if the Rook operator is running in the namespace "my-namespace" the provisioner value should be "my-namespace.rbd.csi.ceph.com". The same provisioner name needs to be set in both the storageclass and snapshotclass.

Liveness Sidecar

All CSI pods are deployed with a sidecar container that provides a prometheus metric for tracking if the CSI plugin is alive and running. These metrics are meant to be collected by prometheus but can be accesses through a GET request to a specific node ip. for example curl -X get http://[pod ip]:[liveness-port][liveness-path] 2>/dev/null | grep csi the expected output should be

curl -X GET http://10.109.65.142:9080/metrics 2>/dev/null | grep csi
# HELP csi_liveness Liveness Probe
# TYPE csi_liveness gauge
csi_liveness 1

Check the monitoring doc to see how to integrate CSI liveness and grpc metrics into ceph monitoring.

Dynamically Expand Volume

Prerequisite

  • For filesystem resize to be supported for your Kubernetes cluster, the kubernetes version running in your cluster should be >= v1.15 and for block volume resize support the Kubernetes version should be >= v1.16. Also, ExpandCSIVolumes feature gate has to be enabled for the volume resize functionality to work.

To expand the PVC the controlling StorageClass must have allowVolumeExpansion set to true. csi.storage.k8s.io/controller-expand-secret-name and csi.storage.k8s.io/controller-expand-secret-namespace values set in storageclass. Now expand the PVC by editing the PVC pvc.spec.resource.requests.storage to a higher values than the current size. Once PVC is expanded on backend and same is reflected size is reflected on application mountpoint, the status capacity pvc.status.capacity.storage of PVC will be updated to new size.

RBD Mirroring

To support RBD Mirroring, the Volume Replication Operator will be started in the RBD provisioner pod. The Volume Replication Operator is a kubernetes operator that provides common and reusable APIs for storage disaster recovery. It is based on csi-addons/spec specification and can be used by any storage provider. It follows the controller pattern and provides extended APIs for storage disaster recovery. The extended APIs are provided via Custom Resource Definitions (CRDs).

Enable volume replication

  1. Install the volume replication CRDs:
kubectl create -f https://raw.githubusercontent.com/csi-addons/volume-replication-operator/v0.1.0/config/crd/bases/replication.storage.openshift.io_volumereplications.yaml
kubectl create -f https://raw.githubusercontent.com/csi-addons/volume-replication-operator/v0.1.0/config/crd/bases/replication.storage.openshift.io_volumereplicationclasses.yaml
  1. Enable the volume replication controller:

generic ephemeral volume support

The generic ephemeral volume feature adds support for specifying PVCs in the volumes field to indicate a user would like to create a Volume as part of the pod spec. This feature requires the GenericEphemeralVolume feature gate to be enabled. Because this is a beta feature, it is enabled by default in kubernetes versions >=1.21.

For example:

kind: Pod
apiVersion: v1
metadata:
  name: csi-rbd-demo-ephemeral-pod
spec:
  containers:
    - name: web-server
      image: docker.io/library/nginx:latest
      volumeMounts:
        - mountPath: "/myspace"
          name: mypvc
  volumes:
    - name: mypvc
      ephemeral:
        volumeClaimTemplate:
          spec:
            accessModes: ["ReadWriteOnce"]
            storageClassName: "rook-ceph-block"
            resources:
              requests:
                storage: 1Gi

A volume claim template is defined inside the pod spec which refers to a volume provisioned and used by the pod with its lifecycle. The volumes are provisioned when pod get spawned and destroyed at time of pod delete.

Refer to ephemeral-doc for more info.

Prerequisites

  1. Requires Kubernetes v1.21+ which supports ephemeral volume.
  2. In [pod-ephemeral](https://github.com/rook/rook/tree/{{ branchName }}/cluster/examples/kubernetes/ceph/csi/rbd/pod-ephemeral.yaml), storageclass should be the name of the storageclass which is already available for the RBD provisioning to work.