diff options
author | OpenShift Merge Robot <openshift-merge-robot@users.noreply.github.com> | 2018-01-11 11:39:02 -0800 |
---|---|---|
committer | GitHub <noreply@github.com> | 2018-01-11 11:39:02 -0800 |
commit | fdc5829d6ca252e1574278cd1dc50e932378d98d (patch) | |
tree | 0f80b64e32b59e7c4a2735066aed5f4e363e5a8a /playbooks/openstack | |
parent | e347c3ed3cc23cdf3af1077dd4becc572a7dfad9 (diff) | |
parent | 3d943161cedc2fc0d1690e03ece1e2504d4d6d74 (diff) | |
download | openshift-fdc5829d6ca252e1574278cd1dc50e932378d98d.tar.gz openshift-fdc5829d6ca252e1574278cd1dc50e932378d98d.tar.bz2 openshift-fdc5829d6ca252e1574278cd1dc50e932378d98d.tar.xz openshift-fdc5829d6ca252e1574278cd1dc50e932378d98d.zip |
Merge pull request #6607 from tomassedovic/fix-cinder-pv
Automatic merge from submit-queue.
Fix Cinder Persistent Volume support
This documents how to use Cinder-backed persistent volumes with OpenStack.
It needed a change to the dynamic inventory because the "openstack" cloudprovider plugin does actually require internal name resolution -- and the `openshift_hostname` value must match the name of the Nova server.
In addition, we need to be able to specify the V2 of the Cinder API for now as described in: https://github.com/openshift/openshift-docs/issues/5730
Diffstat (limited to 'playbooks/openstack')
-rw-r--r-- | playbooks/openstack/advanced-configuration.md | 106 | ||||
-rw-r--r-- | playbooks/openstack/sample-inventory/group_vars/OSEv3.yml | 1 | ||||
-rwxr-xr-x | playbooks/openstack/sample-inventory/inventory.py | 6 |
3 files changed, 111 insertions, 2 deletions
diff --git a/playbooks/openstack/advanced-configuration.md b/playbooks/openstack/advanced-configuration.md index 2c9b70b5f..2caf63592 100644 --- a/playbooks/openstack/advanced-configuration.md +++ b/playbooks/openstack/advanced-configuration.md @@ -372,6 +372,112 @@ In order to set a custom entrypoint, update `openshift_master_cluster_public_hos Note than an empty hostname does not work, so if your domain is `openshift.example.com`, you cannot set this value to simply `openshift.example.com`. + +## Using Cinder-backed Persistent Volumes + +You will need to set up OpenStack credentials. You can try putting this in your +`inventory/group_vars/OSEv3.yml`: + + openshift_cloudprovider_kind: openstack + openshift_cloudprovider_openstack_auth_url: "{{ lookup('env','OS_AUTH_URL') }}" + openshift_cloudprovider_openstack_username: "{{ lookup('env','OS_USERNAME') }}" + openshift_cloudprovider_openstack_password: "{{ lookup('env','OS_PASSWORD') }}" + openshift_cloudprovider_openstack_tenant_name: "{{ lookup('env','OS_PROJECT_NAME') }}" + openshift_cloudprovider_openstack_domain_name: "{{ lookup('env','OS_USER_DOMAIN_NAME') }}" + openshift_cloudprovider_openstack_blockstorage_version: v2 + +**NOTE**: you must specify the Block Storage version as v2, because OpenShift +does not support the v3 API yet and the version detection is currently not +working properly. + +For more information, consult the [Configuring for OpenStack page in the OpenShift documentation][openstack-credentials]. + +[openstack-credentials]: https://docs.openshift.org/latest/install_config/configuring_openstack.html#install-config-configuring-openstack + +**NOTE** the OpenStack integration currently requires DNS to be configured and +running and the `openshift_hostname` variable must match the Nova server name +for each node. The cluster deployment will fail without it. If you use the +provided OpenStack dynamic inventory and configure the +`openshift_openstack_dns_nameservers` Ansible variable, this will be handled +for you. + +After a successful deployment, the cluster is configured for Cinder persistent +volumes. + +### Validation + +1. Log in and create a new project (with `oc login` and `oc new-project`) +2. Create a file called `cinder-claim.yaml` with the following contents: + +```yaml +apiVersion: "v1" +kind: "PersistentVolumeClaim" +metadata: + name: "claim1" +spec: + accessModes: + - "ReadWriteOnce" + resources: + requests: + storage: "1Gi" +``` +3. Run `oc create -f cinder-claim.yaml` to create the Persistent Volume Claim object in OpenShift +4. Run `oc describe pvc claim1` to verify that the claim was created and its Status is `Bound` +5. Run `openstack volume list` + * A new volume called `kubernetes-dynamic-pvc-UUID` should be created + * Its size should be `1` + * It should not be attached to any server +6. Create a file called `mysql-pod.yaml` with the following contents: + +```yaml +apiVersion: v1 +kind: Pod +metadata: + name: mysql + labels: + name: mysql +spec: + containers: + - resources: + limits : + cpu: 0.5 + image: openshift/mysql-55-centos7 + name: mysql + env: + - name: MYSQL_ROOT_PASSWORD + value: yourpassword + - name: MYSQL_USER + value: wp_user + - name: MYSQL_PASSWORD + value: wp_pass + - name: MYSQL_DATABASE + value: wp_db + ports: + - containerPort: 3306 + name: mysql + volumeMounts: + - name: mysql-persistent-storage + mountPath: /var/lib/mysql/data + volumes: + - name: mysql-persistent-storage + persistentVolumeClaim: + claimName: claim1 +``` + +7. Run `oc create -f mysql-pod.yaml` to create the pod +8. Run `oc describe pod mysql` + * Its events should show that the pod has successfully attached the volume above + * It should show no errors + * `openstack volume list` should show the volume attached to an OpenShift app node + * NOTE: this can take several seconds +9. After a while, `oc get pod` should show the `mysql` pod as running +10. Run `oc delete pod mysql` to remove the pod + * The Cinder volume should no longer be attached +11. Run `oc delete pvc claim1` to remove the volume claim + * The Cinder volume should be deleted + + + ## Creating and using a Cinder volume for the OpenShift registry You can optionally have the playbooks create a Cinder volume and set diff --git a/playbooks/openstack/sample-inventory/group_vars/OSEv3.yml b/playbooks/openstack/sample-inventory/group_vars/OSEv3.yml index 481807dc9..a8663f946 100644 --- a/playbooks/openstack/sample-inventory/group_vars/OSEv3.yml +++ b/playbooks/openstack/sample-inventory/group_vars/OSEv3.yml @@ -20,6 +20,7 @@ openshift_hosted_registry_wait: True #openshift_cloudprovider_openstack_password: "{{ lookup('env','OS_PASSWORD') }}" #openshift_cloudprovider_openstack_tenant_name: "{{ lookup('env','OS_TENANT_NAME') }}" #openshift_cloudprovider_openstack_region: "{{ lookup('env', 'OS_REGION_NAME') }}" +#openshift_cloudprovider_openstack_blockstorage_version: v2 ## Use Cinder volume for Openshift registry: diff --git a/playbooks/openstack/sample-inventory/inventory.py b/playbooks/openstack/sample-inventory/inventory.py index 45cc4e15a..76e658eb7 100755 --- a/playbooks/openstack/sample-inventory/inventory.py +++ b/playbooks/openstack/sample-inventory/inventory.py @@ -89,13 +89,15 @@ def build_inventory(): # TODO(shadower): what about multiple networks? if server.private_v4: hostvars['private_v4'] = server.private_v4 + hostvars['openshift_ip'] = server.private_v4 + # NOTE(shadower): Yes, we set both hostname and IP to the private # IP address for each node. OpenStack doesn't resolve nodes by # name at all, so using a hostname here would require an internal # DNS which would complicate the setup and potentially introduce # performance issues. - hostvars['openshift_ip'] = server.private_v4 - hostvars['openshift_hostname'] = server.private_v4 + hostvars['openshift_hostname'] = server.metadata.get( + 'openshift_hostname', server.private_v4) hostvars['openshift_public_hostname'] = server.name if server.metadata['host-type'] == 'cns': |