|
| 1 | +# Using the LVM-Operator |
| 2 | + |
| 3 | +Because there currently is no CI pipeline that builds this repo, you will either have to build it yourself, or use a prebuilt state made available. The prebuilt state might be out of sync with what is currently in the repo |
| 4 | + |
| 5 | +## Preparations |
| 6 | + |
| 7 | +### Building the Operator yourself |
| 8 | + |
| 9 | +Building the operator is easy. Just make sure you have docker or podman installed on your system and that you are logged into your registry (quay.io, docker, ...) |
| 10 | + |
| 11 | +Set the environment variable `IMG` to the new repository path where you want to host your image: |
| 12 | + |
| 13 | +``` |
| 14 | +export IMG=quay.io/USER/lvm-operator |
| 15 | +``` |
| 16 | + |
| 17 | +Then start the build process like this: |
| 18 | + |
| 19 | +``` |
| 20 | +make docker-build-combined docker-push |
| 21 | +``` |
| 22 | + |
| 23 | +Ensure that the new repository in your registry is either set to public or that your target OpenShift cluster has read access to that repository. |
| 24 | + |
| 25 | +When this is finished, you are ready to continue with the deploy steps |
| 26 | + |
| 27 | +### Using the pre-built image |
| 28 | + |
| 29 | +If you are ok with using the prebuilt images, then just set your variable like this: |
| 30 | + |
| 31 | +``` |
| 32 | +export IMG=quay.io/mulbc/lvm-operator |
| 33 | +``` |
| 34 | + |
| 35 | +## Deploy |
| 36 | + |
| 37 | +Ensured that your `IMG` variable is set to a repository that contains the operator. |
| 38 | +Afterwards ensure that you are connected to the right cluster: |
| 39 | + |
| 40 | +``` |
| 41 | +oc get nodes |
| 42 | +``` |
| 43 | + |
| 44 | +After you have ensured both, you can start the deployment with |
| 45 | + |
| 46 | +``` |
| 47 | +make deploy |
| 48 | +``` |
| 49 | + |
| 50 | +After this has finished successfully, you should switch over to the lvm-operator namespace: |
| 51 | + |
| 52 | +`oc project lvm-operator-system` |
| 53 | + |
| 54 | +Please wait until all Pods are finished running: |
| 55 | + |
| 56 | +``` |
| 57 | +oc get pods -w |
| 58 | +``` |
| 59 | + |
| 60 | +After the controller is running, create the sample lvmCluster CR: |
| 61 | + |
| 62 | +``` |
| 63 | +oc create -f https://github.com/red-hat-storage/lvm-operator/raw/main/config/samples/lvm_v1alpha1_lvmcluster.yaml |
| 64 | +``` |
| 65 | + |
| 66 | +This will try to leverage all nodes and all available disks on these nodes. |
| 67 | + |
| 68 | +Again please wait until all Pods are finished running: |
| 69 | + |
| 70 | +``` |
| 71 | +oc get pods -w |
| 72 | +``` |
| 73 | + |
| 74 | +The `topolvm-node` pod will be in init until `vg-manager` has done all the preparation and this might take a while. |
| 75 | + |
| 76 | +After all Pods are running, you will get a storage class that you can use when creating a PVCs: |
| 77 | + |
| 78 | +``` |
| 79 | +oc get sc |
| 80 | +
|
| 81 | +NAME PROVISIONER RECLAIMPOLICY VOLUMEBINDINGMODE ALLOWVOLUMEEXPANSION AGE |
| 82 | +topolvm-vg1 topolvm.cybozu.com Delete WaitForFirstConsumer true 3m44s |
| 83 | +``` |
| 84 | + |
| 85 | +## Test |
| 86 | + |
| 87 | +After the operator is installed and the Cluster is set up, you can start by creating a simple test application that will consume storage: |
| 88 | + |
| 89 | +Create the PVC: |
| 90 | + |
| 91 | +```yaml |
| 92 | +cat <<EOF | oc apply -f - |
| 93 | +apiVersion: v1 |
| 94 | +kind: PersistentVolumeClaim |
| 95 | +metadata: |
| 96 | + name: lvmpvc |
| 97 | + labels: |
| 98 | + type: local |
| 99 | +spec: |
| 100 | + storageClassName: topolvm-vg1 |
| 101 | + resources: |
| 102 | + requests: |
| 103 | + storage: 5Gi |
| 104 | + accessModes: |
| 105 | + - ReadWriteOnce |
| 106 | + volumeMode: Filesystem |
| 107 | +EOF |
| 108 | +``` |
| 109 | + |
| 110 | +You will see that the PVC will be stuck in the pending state: |
| 111 | + |
| 112 | +``` |
| 113 | +oc get pvc |
| 114 | +
|
| 115 | +NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE |
| 116 | +lvmpvc Pending topolvm-vg1 7s |
| 117 | +``` |
| 118 | + |
| 119 | +That's because our Storage Class is waiting for a Pod that needs that PVC, before it is created. So let's create a Pod for this PVC. |
| 120 | + |
| 121 | +```yaml |
| 122 | +cat <<EOF | oc apply -f - |
| 123 | +apiVersion: v1 |
| 124 | +kind: Pod |
| 125 | +metadata: |
| 126 | + name: lvmpod |
| 127 | +spec: |
| 128 | + volumes: |
| 129 | + - name: storage |
| 130 | + persistentVolumeClaim: |
| 131 | + claimName: lvmpvc |
| 132 | + containers: |
| 133 | + - name: container |
| 134 | + image: nginx |
| 135 | + ports: |
| 136 | + - containerPort: 80 |
| 137 | + name: "http-server" |
| 138 | + volumeMounts: |
| 139 | + - mountPath: "/usr/share/nginx/html" |
| 140 | + name: storage |
| 141 | +EOF |
| 142 | +``` |
| 143 | + |
| 144 | +After we added the Pod, the PVC will be bound and the Pod will eventually be Running: |
| 145 | + |
| 146 | +``` |
| 147 | +oc get pvc,pods |
| 148 | +
|
| 149 | +NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE |
| 150 | +persistentvolumeclaim/lvmpvc Bound pvc-a37ef71c-a9b9-45d8-96e8-3b5ad30a84f6 5Gi RWO topolvm-vg1 3m2s |
| 151 | +
|
| 152 | +NAME READY STATUS RESTARTS AGE |
| 153 | +pod/lvmpod 1/1 Running 0 28s |
| 154 | +``` |
0 commit comments