Storage

Hosted Cluster storage

Using the UI create a PVC inside of the guest cluster (Ensure you are in the cluster cluster1)

Navigate to Storage → PersistentVolumeClaims
Press Create PersistentVolumeClaim on the top-right
Fill with the name pvc-1 and the size 1 GiB
The PVC will be created and switch to Bound status almost immediately.
Now in the main cluster, navigate to Storage → PersistentVolumeClaims and select project clusters-cluster1. Sort by Capacity

You will see the PVC that has been created for use by the hosted cluster.

Etcd storage

Because etcd writes data to disk and persists proposals on disk, its performance depends on disk performance. Although etcd is not particularly I/O intensive, it requires a low latency block device for optimal performance and stability.

In terms of latency, run etcd on top of a block device that can write at least 50 IOPS of 8000 bytes long sequentially.
That is, with a latency of 10ms, keep in mind that uses fdatasync to synchronize each write in the WAL.
For heavy loaded clusters, sequential 500 IOPS of 8000 bytes (2 ms) are recommended.
To achieve such performance, run etcd on machines that are backed by SSD or NVMe disks with low latency and high throughput.

Review the current etcd PVCs for the cluster1

Later in this lab a new cluster is created using local disks for etcd.

OpenShift Virtualization image caching

By default a CoreOS image is downloaded for each of the workers, which will cause more storage used and more overhead to import the image.

Image caching is an advanced feature that you can use to optimize both cluster startup time and storage utilization.

This feature requires the use of a storage class that is capable of smart cloning and the ReadWriteMany access mode.

Image caching works as follows:

The VM image is imported to a PVC that is associated with the hosted cluster.
A unique clone of that PVC is created for every KubeVirt VM that is added as a worker node to the cluster.

Image caching reduces VM startup time by requiring only a single image import. It can further reduce overall cluster storage usage when the storage class supports copy-on-write cloning.

Destroy current cluster

Delete the managed cluster resource on multicluster engine operator by running the following command:
```
oc delete managedcluster cluster1
```
Sample Output
```
managedcluster.cluster.open-cluster-management.io "cluster1" deleted
```
This command takes a while to return - be patient and wait until you’re back at the prompt.

Delete the hosted cluster and its back-end resources by running the following command

hcp destroy cluster kubevirt --name cluster1

Sample Output

2024-01-02T21:09:35Z    INFO    Found hosted cluster    {"namespace": "clusters", "name": "cluster1"}
2024-01-02T21:09:36Z    INFO    Updated finalizer for hosted cluster    {"namespace": "clusters", "name": "cluster1"}
2024-01-02T21:09:36Z    INFO    Deleting hosted cluster {"namespace": "clusters", "name": "cluster1"}
2024-01-02T21:11:24Z    INFO    Deleting Secrets        {"namespace": "clusters"}
2024-01-02T21:11:24Z    INFO    Deleted CLI generated secrets
2024-01-02T21:11:24Z    INFO    Finalized hosted cluster        {"namespace": "clusters", "name": "cluster1"}
2024-01-02T21:11:24Z    INFO    Successfully destroyed cluster and infrastructure       {"namespace": "clusters", "name": "cluster1", "infraID": "cluster1-qwpqk"}

Again this command takes a while to return - be patient and wait until you’re back at the prompt.

Create a new hosted cluster using image caching and local storage for etcd

The hosting cluster is already configured to use LVM Storage to provide local storage.

Navigate to Operators → Installed Operators
Search for LVM Storage
Click to get the details and navigate to tab LVMCluster

Click on my-lvmcluster and review the YAML

<<REDACTED>>
spec:
  storage:
    deviceClasses:
      - deviceSelector:
          paths:
            - /dev/vdd
        fstype: xfs
        name: vg1
        thinPoolConfig:
          name: thin-pool-1
          overprovisionRatio: 10
          sizePercent: 90
<<REDACTED>>

Navigate to Storage → StorageClasses. A storageclass named lvms-<name> is created automatically when a LVMCluster is configured. In this cluster it is called lvms-vg1

Create a new hosted cluster named cluster2 using the hcp command:

hcp create cluster kubevirt \
--name cluster2 \
--release-image quay.io/openshift-release-dev/ocp-release:4.14.9-x86_64 \
--node-pool-replicas 2 \
--pull-secret ~/pull-secret.json \
--memory 6Gi \
--cores 2 \
--root-volume-cache-strategy=PVC \
--etcd-storage-class lvms-vg1

While the new cluster is being created, review the Persistent Volume Claims in the namespace clusters-cluster2
Notice a image with prefix kv-boot-image-cache is created.
Select one of the PVCs starting with cluster2 and scroll down to find the Owner.

It may take a few minutes for those PVCs to appear. Be patient and wait until you see two PVCs whose name starts with cluster2.
Information about the DataVolume will appear, scrolling down you can see information about the clone process.