ceph-csi/docs/static-pvc.md
Madhu Rajanna 200fbdbf28 doc: fix fsName details in PV object
fsName is optional and its required
to mount the filesystem incase if we have
multiple filesystem in the cluster.

Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
2024-01-12 18:27:45 +00:00

10 KiB

Static PVC with ceph-csi

This document outlines how to create static PV and static PVC from existing rbd image/cephFS volume.

warning static PVC can be created, deleted, mounted and unmounted but currently ceph-csi doesn't support other operations like snapshot,clone, resize, etc for static PVC

RBD static PVC

RBD images created manually can be mounted and unmounted to an app, below step shows how to create a rbd image, static PV, static PVC

Create RBD image

If you already have a rbd image created and contains some data which you want to access by the application pod you can skip this step.

Lets create a new rbd image in ceph cluster which we are going to use for static PVC

rbd create static-image --size=1024 --pool=replicapool

Create RBD static PV

To create the rbd PV you need to know the rbd image name,clusterID and pool name in which the rbd image is created

apiVersion: v1
kind: PersistentVolume
metadata:
  name: fs-static-pv
spec:
  accessModes:
  - ReadWriteOnce
  capacity:
    storage: 1Gi
  csi:
    driver: rbd.csi.ceph.com
    fsType: ext4
    nodeStageSecretRef:
      # node stage secret name
      name: csi-rbd-secret
      # node stage secret namespace where above secret is created
      namespace: default
    volumeAttributes:
      # Required options from storageclass parameters need to be added in volumeAttributes
      "clusterID": "ba68226a-672f-4ba5-97bc-22840318b2ec"
      "pool": "replicapool"
      "staticVolume": "true"
      "imageFeatures": "layering"
      #mounter: rbd-nbd
    # volumeHandle should be same as rbd image name
    volumeHandle: static-image
  persistentVolumeReclaimPolicy: Retain
  # The volumeMode can be either `Filesystem` or `Block` if you are creating Filesystem PVC it should be `Filesystem`, if you are creating Block PV you need to change it to `Block`
  volumeMode: Filesystem

RBD Volume Attributes in PV

Below table explains the list of volume attributes can be set when creating a static RBD PV

Attributes Description Required
clusterID The clusterID is used by the CSI plugin to uniquely identify and use a Ceph cluster (this is the key in configmap created duing ceph-csi deployment) Yes
pool The pool name in which rbd image is created Yes
staticVolume Value must be set to true to mount and unmount static rbd PVC Yes
imageFeatures CSI RBD currently supports layering, journaling, exclusive-lock features. If journaling is enabled, must enable exclusive-lock too Yes
mounter If set to rbd-nbd, use rbd-nbd on nodes that have rbd-nbd and nbd kernel modules to map rbd images No

Note ceph-csi does not supports rbd image deletion for static PV. persistentVolumeReclaimPolicy in PV spec must be set to Retain to avoid PV delete attempt in csi-provisioner.

$ kubectl create -f fs-static-pv.yaml
persistentvolume/fs-static-pv created

Create RBD static PVC

To create the rbd PVC you need to know the PV name which is created above

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: fs-static-pvc
  namespace: default
spec:
  storageClassName: ""
  accessModes:
  # ReadWriteMany is only supported for Block PVC
  - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
  # The volumeMode can be either `Filesystem` or `Block` if you are creating Filesystem PVC it should be `Filesystem`, if you are creating Block PV you need to change it to `Block`
  volumeMode: Filesystem
  # volumeName should be same as PV name
  volumeName: fs-static-pv
$ kubectl create -f fs-static-pvc.yaml
persistentvolumeclaim/fs-static-pvc created

Resize RBD image

Let us resize the RBD image in ceph cluster

rbd resize static-image --size=2048 --pool=replicapool

Once the rbd image is resized in the ceph cluster, update the PV size and PVC size to match the size of the rbd image.

Now scale down the application pod which is using cephfs-static-pvc and scale up the application pod to resize the filesystem.

Note If you have mounted same static PVC to multiple application pods, make sure you will scale down all the application pods and make sure no application pods using the static PVC is running on the node and scale up all the application pods again(this will trigger NodeStageVolumeRequest which will resize the filesystem for static volume).

Note deleting PV and PVC does not removed the backend rbd image, user need to manually delete the rbd image if required

CephFS static PVC

CephFS subvolume created manually can be mounted and unmounted to an app, below steps show how to create a CephFS subvolume, static PV and static PVC.

Create CephFS subvolume

If you already have a CephFS subvolume created and contains some data which you want to access by the application pod you can skip this step.

Lets create a new CephFS subvolume of size 1 GiB in ceph cluster which we are going to use for static PVC, before that we need to create the subvolumegroup. myfs is the filesystem name(volume name) inside which subvolume should be created.

ceph fs subvolumegroup create myfs testGroup
ceph fs subvolume create myfs testSubVolume testGroup --size=1073741824

Note: volume here refers to the filesystem.

Create CephFS static PV

To create the CephFS PV you need to know the volume rootpath, and clusterID, here is the command to get the root path in ceph cluster

$ ceph fs subvolume getpath myfs testSubVolume testGroup
/volumes/testGroup/testSubVolume
apiVersion: v1
kind: PersistentVolume
metadata:
  name: cephfs-static-pv
spec:
  accessModes:
  - ReadWriteMany
  capacity:
    storage: 1Gi
  csi:
    driver: cephfs.csi.ceph.com
    nodeStageSecretRef:
      # node stage secret name
      name: csi-cephfs-secret
      # node stage secret namespace where above secret is created
      namespace: default
    volumeAttributes:
      # optional file system to be mounted
      "fsName": "myfs"
      # Required options from storageclass parameters need to be added in volumeAttributes
      "clusterID": "ba68226a-672f-4ba5-97bc-22840318b2ec"
      "staticVolume": "true"
      "rootPath": /volumes/testGroup/testSubVolume
    # volumeHandle can be anything, need not to be same
    # as PV name or volume name. keeping same for brevity
    volumeHandle: cephfs-static-pv
  persistentVolumeReclaimPolicy: Retain
  volumeMode: Filesystem

Node stage secret ref in CephFS PV

For static CephFS PV to work, userID and userKey needs to be specified in the secret. Static PV will not work with adminID and adminKey. Format for the secret should be same as detailed here.

CephFS volume attributes in PV

Below table explains the list of volume attributes can be set when creating a static CephFS PV

Attributes Description Required
clusterID The clusterID is used by the CSI plugin to uniquely identify and use a Ceph cluster (this is the key in configmap created duing ceph-csi deployment) Yes
fsName CephFS filesystem name to be mounted. Not passing this option mounts the default file system. No
staticVolume Value must be set to true to mount and unmount static cephFS PVC Yes
rootPath Actual path of the subvolume in ceph cluster, can be retrieved by issuing getpath command as described above Yes

Note ceph-csi does not supports CephFS subvolume deletion for static PV. persistentVolumeReclaimPolicy in PV spec must be set to Retain to avoid PV delete attempt in csi-provisioner.

$ kubectl create -f cephfs-static-pv.yaml
persistentvolume/cephfs-static-pv created

Create CephFS static PVC

To create the CephFS PVC you need to know the PV name which is created above

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: cephfs-static-pvc
  namespace: default
spec:
  accessModes:
  - ReadWriteMany
  resources:
    requests:
      storage: 1Gi
  storageClassName: ""
  volumeMode: Filesystem
  # volumeName should be same as PV name
  volumeName: cephfs-static-pv
$ kubectl create -f cephfs-static-pvc.yaml
persistentvolumeclaim/cephfs-static-pvc created

Note deleting PV and PVC does not delete the backend CephFS subvolume, user needs to manually delete the CephFS subvolume if required.