Skip to content

Latest commit

 

History

History
167 lines (136 loc) · 6.08 KB

upload.md

File metadata and controls

167 lines (136 loc) · 6.08 KB
Raw

CDI Upload User Guide

The purpose of this document is to show how to upload a VM disk image on your local system to a PersistentVolumeClaim in Kubernetes.

Prerequesites

You have a Kubernetes cluster up and running with CDI installed and at least one PersistentVolume is available.

Commands/manifests below will be run from the root of the CDI repo against a Minikube cluster.

If you are using Minikube with the storage-provisioner addon enabled. You can create a PersistentVolume like so:

cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: PersistentVolume
metadata:
  name: pv0001
spec:
  accessModes:
    - ReadWriteOnce
  capacity:
    storage: 5Gi
  hostPath:
    path: /data/pv0001/
EOF

Expose cdi-uploadproxy service

In order to upload data to your cluster, the cdi-uploadproxy service must be accessible from outside the cluster. In a production environment, this probably involves setting up a Ingress or a LoadBalancer Service.

Minikube

cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Service
metadata:
  name: cdi-uploadproxy-nodeport
  namespace: cdi
  labels:
    cdi.kubevirt.io: "cdi-uploadproxy"
spec:
  type: NodePort
  ports:
    - port: 443
      targetPort: 8443
      nodePort: 31001
      protocol: TCP
  selector:
    cdi.kubevirt.io: cdi-uploadproxy
EOF

Minishift

oc get cm -n cdi cdi-uploadproxy-signer-bundle -o=jsonpath="{.data['ca-bundle\.crt']}" > tls.crt && \
oc create route reencrypt -n cdi --service=cdi-uploadproxy --dest-ca-cert=tls.crt && \
rm tls.crt

Port forwarding via the API server

kubectl port-forward -n cdi service/cdi-uploadproxy 8443:443

(Make sure port 8443 on your system isn't occupied.)

Create a Data Volume

Specifying an 'upload' source will mark the data volume as a target for upload.

To create an upload datavolume use the following example.

apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: upload-datavolume
spec:
  source:
      upload: {}
  storage:
    resources:
      requests:
        storage: 500Mi
kubectl apply -f manifests/example/upload-datavolume.yaml

Create a Data Volume for archive upload

You can also upload an archive. Specifying in the data volume spec: contentType: archive will mark the datavolume as archive upload and will handle the content as needed (supports also compressed tar)

Request an Upload Token

Before sending data to the Upload Proxy, an Upload Token must be requested.

Take a look at at manifests/example/upload-datavolume-token.yaml for an example.

apiVersion: upload.cdi.kubevirt.io/v1beta1
kind: UploadTokenRequest
metadata:
  name: upload-datavolume
  namespace: default
spec:
  pvcName: upload-datavolume

kubectl apply -f manifests/example/upload-datavolume-token.yaml -o yaml
apiVersion: upload.cdi.kubevirt.io/v1beta1
kind: UploadTokenRequest
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"upload.cdi.kubevirt.io/v1beta1","kind":"UploadTokenRequest","metadata":{"annotations":{},"name":"upload-datavolume-token","namespace":"default"},"spec":{"pvcName":"upload-datavolume"}}
  creationTimestamp: null
  name: upload-datavolume-token
  namespace: default
spec:
  pvcName: upload-datavolume
status:
  token: eyJhbGciOiJQUzUxMiIsImtpZCI6IiJ9.eyJwdmNOYW1lIjoidXBsb2FkLXRlc3QiLCJuYW1lc3BhY2UiOiJkZWZhdWx0IiwiY3JlYXRpb25UaW1lc3RhbXAiOiIyMDE4LTA5LTIxVDE4OjEyOjE5LjQwODI1MDQ4NFoifQ.JWk1VyvzSse3eFiBROKgGoLnOPCiYW9JdDWKXFROEL6XY0O5lFb1R0rwdfWwC3BBOtEA9mC9x3ZGYPnYWO-5G_r1fWKHjF-zifrCX_3Dhp3vfSq6Zfpu-vV0Qn0A3YkSCCmiC_nONAhVjEDuQsRFIKwYcxBoEOpye92ggH2u5FxQE7FwxxH6-RHun9tc_lIFX-ZFKnq7n5tWbjsTmAZI_4rDNgYkVFhFtENU6e-5_Ncokxs3YVzkbSrXweZpRmmaYQOmZhjXSLjKED_2FVq7tYeVueEEhKC_zJ-AEivstALPwPjiwyWXJyfE3dCmbA1sBKuNUrAaDlBvSAp1uPV9eQ

Save the token field of the response status. It will be used to authorize our CDI Upload request. Tokens are good for 5 minutes.

You can capture the token in an environment variable by doing this:

TOKEN=$(kubectl apply -f manifests/example/upload-datavolume-token.yaml -o="jsonpath={.status.token}")

Upload an Image

We will be using curl to upload tests/images/cirros-qcow2.img to the datavolume.

Assuming that the environment variable TOKEN contains a valid UploadToken, execute the following to upload the image:

Minikube

Synchronous

curl -v --insecure -H "Authorization: Bearer $TOKEN" --data-binary @tests/images/cirros-qcow2.img https://$(minikube ip):31001/v1beta1/upload

The connection will not be closed until the entire process is completed. If the conversion or resizing process takes a long time intermediate proxies might close the connection unexpectedly.

Asynchronous

curl -v --insecure -H "Authorization: Bearer $TOKEN" --data-binary @tests/images/cirros-qcow2.img https://$(minikube ip):31001/v1beta1/upload-async

As soon as the data has been transmitted, the connection will be closed. The caller should monitor the Datavolume status to see if the process is completed.

Minishift

Synchronous

curl -v --insecure -H "Authorization: Bearer $TOKEN" --data-binary @tests/images/cirros-qcow2.img https://cdi-uploadproxy-cdi.$(minishift ip).nip.io/v1beta1/upload

The connection will not be closed until the entire process is completed. If the conversion or resizing process takes a long time intermediate proxies might close the connection unexpectedly.

Asynchronous

curl -v --insecure -H "Authorization: Bearer $TOKEN" --data-binary @tests/images/cirros-qcow2.img https://cdi-uploadproxy-cdi.$(minishift ip).nip.io/v1beta1/upload-async

As soon as the data has been transmitted, the connection will be closed. The caller should monitor the Datavolume status to see if the process is completed.

Assuming you did not get an error, the Datavolume upload-datavolume should now contain a bootable VM image.

Using Kubevirt image upload

If you have also Kubevirt extension you can use virtctl image-upload. For examples check out image-upload help.