Installing Red Hat CloudForms on OpenShift Container Platform

Red Hat CloudForms 4.6

How to install and configure Red Hat CloudForms on an OpenShift Container Platform environment

Red Hat CloudForms Documentation Team

Abstract

This guide provides instructions on how to deploy and configure a Red Hat CloudForms appliance as multiple pods on an OpenShift Container Platform environment.
If you have a suggestion for improving this guide or have found an error, please submit a Bugzilla report at http://bugzilla.redhat.com against Red Hat CloudForms Management Engine for the Documentation component. Please provide specific details, such as the section number, guide name, and CloudForms version so we can easily locate the content.

Chapter 1. Installing Red Hat CloudForms

Red Hat CloudForms can be installed on OpenShift Container Platform in a few steps.

This procedure uses a template to deploy a multi-pod CloudForms appliance with the database stored in a persistent volume on OpenShift Container Platform. It provides a step-by-step setup, including cluster administrative tasks as well as information and commands for the application developer using the deployment.

The ultimate goal of the deployment is to be able to deconstruct the CloudForms appliance into several containers running on a pod or a series of pods.

Running the CloudForms appliance in a series of pods has several advantages. For example, running each worker in a separate pod allows OpenShift Container Platform to manage worker processes and reduce worker memory consumption. OpenShift can also easily scale workers by adding or removing pods, and perform upgrades by using images.

There are two options for installing CloudForms on OpenShift:

  • During OpenShift Container Platform 3.7 installation:

    • When you install OpenShift Container Platform 3.7, you have the option to install CloudForms inside OpenShift at the time. This method leverages the Ansible installer to run and deploy the CloudForms template, instead of building the environment manually. See the OpenShift Container Platform 3.7 Release Notes for details.
  • Manual install on an existing OpenShift Container Platform environment:

    • Deploy CloudForms pods using the CloudForms template (.yaml file). This is the method described in this guide.

After deployment, you can configure the CloudForms environment to use any external authentication configurations supported by CloudForms.

1.1. Prerequisites

To successfully deploy a CloudForms appliance on OpenShift Container Platform, you need a functioning OpenShift Container Platform 3.7 install with the following configured:

  • NFS or other compatible volume provider
  • A cluster-admin user
  • A regular user (such as an application developer)
Important

OpenShift Container Platform 3.7 is required for this installation. Red Hat has not tested this procedure with earlier versions of OpenShift Container Platform.

1.1.1. Cluster Sizing

To avoid deployment failures due to resource starvation, Red Hat recommends the following minimum cluster size for a test environment:

  • 1 master node with at least 8 vCPUs and 12GB RAM
  • 2 schedulable nodes with at least 4 vCPUs and 8GB RAM
  • 25GB storage for CloudForms physical volume use

These recommendations assume CloudForms is the only application running on this cluster. Alternatively, you can provision an infrastructure node to run registry, metrics, router, and logging pods.

Each CloudForms application pod will consume at least 3GB RAM on initial deployment (without providers added). RAM consumption increases depending on the appliance use. For example, after adding providers, expect higher resource consumption.

1.1.2. Limitations

The following limitations exist when deploying this version of CloudForms on OpenShift Container Platform 3.7:

  • This configuration cannot run on public OpenShift (OpenShift.io and OpenShift Dedicated environments) because of necessary privileges
  • The Embedded Ansible pod must run as a privileged pod
  • OpenShift cannot independently scale workers
  • A highly available database is not supported in PostgreSQL pods

1.1.3. Templates and Images

The CloudForms deployment uses .yaml template files to create the appliance, including cfme-template.yaml, which is the CloudForms template used for the deployment, and cfme-pv-example.yaml and cfme-pv-app-example.yaml, two pod volume files.

These templates are available in RPMs from Red Hat-provided image streams. To obtain the templates:

  1. Configure image streams as described in OpenShift Container Platform Installation and Configuration.
  2. After loading the image streams and templates, the templates will be available on your OpenShift system in /usr/share/ansible/openshift-ansible/roles/openshift_examples/files/examples/v3.7/cfme-templates.

The CloudForms template points to several image files to create the OpenShift pods that comprise the appliance. These image files are obtained from the Red Hat Container Catalog during deployment.

1.2. Preparing to Deploy CloudForms

To prepare for deploying the CloudForms appliance to OpenShift Container Platform, create a project, configure security contexts, and create persistent storage.

  1. As a regular user, log in to OpenShift:

    $ oc login -u <user> -p <password>
  2. Create a project with your desired parameters. The project name (<your_project> in this example) is mandatory, but <description> and <display_name> are optional:

    $ oc new-project <your_project> \
    --description="<description>" \
    --display-name="<display_name>"
  3. As the admin user, configure security context constraints (SCCs) for your OpenShift service accounts:

    1. Add the cfme-anyuid service account to the anyuid SCC:

      $ oc adm policy add-scc-to-user anyuid system:serviceaccount:<your-project>:cfme-anyuid
    2. Add the cfme-orchestrator service account to the anyuid SCC:

      $ oc adm policy add-scc-to-user anyuid system:serviceaccount:<your-project>:cfme-orchestrator
    3. Add the cfme-httpd service account to the anyuid SCC:

      $ oc adm policy add-scc-to-user anyuid system:serviceaccount:<your-project>:cfme-httpd
    4. Add the cfme-privileged service account to the privileged SCC:

      $ oc adm policy add-scc-to-user privileged system:serviceaccount:<your-project>:cfme-privileged
  4. Verify the SCCs are added correctly to the service accounts and project:

    $ oc describe scc anyuid | grep Users
    Users:					system:serviceaccount:<your-project>:cfme-anyuid,system:serviceaccount:<your-project>:cfme-httpd,system:serviceaccount:<your-project>:cfme-orchestrator
    $ oc describe scc privileged | grep Users
    Users:					system:admin,system:serviceaccount:openshift-infra:build-controller,system:serviceaccount:management-infra:management-admin,system:serviceaccount:management-infra:inspector-admin,system:serviceaccount:logging:aggregated-logging-fluentd,system:serviceaccount:<your-project>:cfme-privileged
    Note

    For more information on SCCs, see the OpenShift documentation.

  5. Add the view and edit roles to the cfme-orchestrator service account:

    $ oc policy add-role-to-user view system:serviceaccount:<your-project>:cfme-orchestrator -n <your-project>
    $ oc policy add-role-to-user edit system:serviceaccount:<your-project>:cfme-orchestrator -n <your-project>
  6. As the admin user, prepare persistent storage for the deployment. (Skip this step if you have already configured persistent storage.)

    A basic CloudForms deployment needs at least two persistent volumes (PVs) to store CloudForms data. As the admin user, create two persistent volumes: one to host the CloudForms PostgreSQL database, and one to host the application data.

    Example NFS-backed volume templates are provided by cfme-pv-db-example.yaml and cfme-pv-server-example.yaml, available from GitHub.

    Note

    For NFS-backed volumes, ensure your NFS server firewall is configured to allow traffic on port 2049 (TCP) from the OpenShift cluster.

    Red Hat recommends setting permissions for the pv-app (privileged pod volume) as 777, uid/gid 0 (owned by root). For more information on configuring persistent storage in OpenShift Container Platform, see the OpenShift Container Platform Installation and Configuration guide.

    1. Configure your NFS server host details within these files, and edit any other settings needed to match your environment.
    2. Run the following commands to create the two persistent volumes:

      $ oc create -f cfme-pv-db-example.yaml
      $ oc create -f cfme-pv-server-example.yaml
    3. Process the templates, editing the NFS_HOST parameter (mandatory) and any other parameters:

      $ oc process cfme-pv-db-example.yaml -p NFS_HOST=nfs.example.com | oc create -f -
      
      $ oc process cfme-pv-server-example.yaml -p NFS_HOST=nfs.example.com | oc create -f -
      Note

      There are three parameters required to process the template. Only NFS_HOST is required, PV_SIZE and BASE_PATH contain defaults that do not need editing unless desired:

      • PV_SIZE - Defaults to the recommended PV size for the App/DB template (5Gi/15Gi respectively)
      • BASE_PATH - Defaults to /exports
      • NFS_HOST - No Default - Hostname or IP address of the NFS server
    4. Verify the persistent volumes were created successfully:

      $ oc get pv
      NAME                CAPACITY   ACCESSMODES   RECLAIMPOLICY   STATUS      CLAIM         STORAGECLASS   REASON    AGE
      cfme-app            5Gi        RWO           Retain          Available                                          16s
      
      cfme-db             15Gi       RWO           Retain          Available                                          49s
      Note

      Red Hat recommends validating NFS share connectivity from an OpenShift node before attempting a deployment.

  7. Increase the maximum number of imported images on ImageStream.

    By default, OpenShift Container Platform can import five tags per image stream, but the CloudForms repositories contain more than five images for deployments.

    You can modify this setting on the master node at /etc/origin/master/master-config.yaml so OpenShift can import additional images.

    1. Add the following at the end of the /etc/origin/master/master-config.yaml file:

      ...
      imagePolicyConfig:
        maxImagesBulkImportedPerRepository: 100
    2. Restart the master service:

      $ systemctl restart atomic-openshift-master
  8. On each OpenShift node, persistently enable the container_manage_cgroup SELinux boolean to allow container processes to make changes to cgroup configuration:

    # setsebool -P container_manage_cgroup on

1.3. Deploying the CloudForms Appliance

To deploy the appliance on OpenShift Container Platform, create the CloudForms template and verify it is available in your project.

  1. As a regular user, create the CloudForms template:

    $ oc create -f cfme-template.yaml
    template "cloudforms" created
  2. Verify the template is available with your project:

    $ oc get templates
    NAME         DESCRIPTION                                    PARAMETERS        OBJECTS
    cloudforms   CloudForms appliance with persistent storage   18 (1 blank)      12
  3. (Optional) Customize the template’s deployment parameters. Use the following command to see the available parameters and descriptions:

    $ oc process --parameters -n <your-project> cloudforms

    To customize the deployment configuration parameters, run:

    $ oc edit dc/<deployconfig_name>
  4. To deploy CloudForms from template using default settings, run:

    $ oc new-app --template=cloudforms

    Alternatively, to deploy CloudForms from a template using customized settings, add the -p option and the desired parameters to the command. For example:

    $ oc new-app --template=cloudforms -p DATABASE_VOLUME_CAPACITY=2Gi,POSTGRESQL_MEM_LIMIT=4Gi,APPLICATION_DOMAIN=hostname
    Important

    The APPLICATION_DOMAIN parameter specifies the hostname used to reach the CloudForms application, which eventually constructs the route to the CloudForms pod. If you do not specify the APPLICATION_DOMAIN parameter, the CloudForms application will not be accessible after the deployment; however, this can be fixed by changing the route. For more information on OpenShift template parameters, see the OpenShift Container Platform Developer Guide.

1.3.1. Deploying the CloudForms Appliance Using an External Database

Before attempting to deploy CloudForms using an external database deployment, ensure the following conditions are satisfied:

  • Your OpenShift cluster can access the external PostgreSQL server
  • The CloudForms user, password, and role have been created on the external PostgreSQL server
  • The intended CloudForms database is created, and ownership has been assigned to the CloudForms user

To deploy the appliance:

  1. Import the CloudForms external database template:

    $ oc create -f templates/cfme-template-ext-db.yaml
  2. Launch the deployment with the following command. The database server IP address is required, and the other settings must match your remote PostgreSQL server.

    $ oc new-app --template=cloudforms-ext-db -p DATABASE_IP=<server_ip> -p DATABASE_USER=<user> -p DATABASE_PASSWORD=<password> -p DATABASE_NAME=<database_name>

1.4. Verifying the Configuration

Verify the deployment was successful by running the following commands as a regular user under the CloudForms project:

Note

The first deployment can take several minutes to complete while OpenShift downloads the necessary images.

  1. Confirm the CloudForms pod is bound to the correct security context constraints:

    1. List and obtain the name of the cfme-app pod:

      $ oc get pod
      NAME                 READY     STATUS    RESTARTS   AGE
      cloudforms-0         1/1       Running   0          4m
      httpd-1-w486v        1/1       Running   0          4m
      memcached-1-4xtjc    1/1       Running   0          4m
      postgresql-1-n5tm6   1/1       Running   0          4m
    2. Export the configuration of the pod:

      $ oc export pod <cfme_pod_name>
    3. Examine the output to verify that openshift.io/scc has the value anyuid:

      ...
      metadata:
        annotations:
          openshift.io/scc: anyuid
      ...
  2. Verify the persistent volumes are attached to the postgresql and cfme-app pods:

    $ oc volume pods --all
    pods/postgresql-1-437jg
      pvc/cfme-pgdb-claim (allocated 2GiB) as cfme-pgdb-volume
        mounted at /var/lib/pgsql/data
      secret/default-token-2se06 as default-token-2se06
        mounted at /var/run/secrets/kubernetes.io/serviceaccount
    pods/cfme-1-s3bnp
      pvc/cfme (allocated 2GiB) as cfme-app-volume
        mounted at /persistent
      secret/default-token-9q4ge as default-token-9q4ge
        mounted at /var/run/secrets/kubernetes.io/serviceaccount
  3. Check the readiness of the CloudForms pod:

    Note

    Allow approximately five minutes once pods are in running state for CloudForms to start responding on HTTPS.

    $ oc describe pods <cfme_pod_name>
    ...
    Conditions:
      Type      Status
      Ready     True
    Volumes:
    ...
  4. After you have successfully validated your CloudForms deployment, disable automatic image change triggers to prevent unintended upgrades.

    By default, on initial deployments the automatic image change trigger is enabled. This could potentially start an unintended upgrade on a deployment if a newer image is found in the ImageStream.

    Disable the automatic image change triggers for CloudForms deployment configurations (DCs) on each project with the following commands:

    $ oc set triggers dc --manual -l app=cloudforms
    deploymentconfig "memcached" updated
    deploymentconfig "postgresql" updated
    
    $ oc set triggers dc --from-config --auto -l app=cloudforms
    deploymentconfig "memcached" updated
    deploymentconfig "postgresql" updated
    Note

    The configuration change trigger is kept enabled; to have full control of your deployments, you can alternatively turn it off. See the OpenShift Container Platform Developer Guide for more information on deployment triggers.

1.5. Logging into CloudForms

As part of the deployment, a route to the CloudForms appliance is created for HTTPS access. Once the pods have been successfully deployed, you can log into CloudForms.

You can obtain the CloudForms host address from the project in the OpenShift user interface, or by opening a shell on the pod and getting the route information.

  1. To open a shell on the pod, run:

    $ oc rsh <pod_name> bash -l
  2. Get the route information:

    $ oc get routes
    NAME         HOST/PORT                   PATH                SERVICE      TERMINATION   LABELS
    cloudforms   cfme.apps.e2e.example.com  cloudforms:443-tcp   passthrough                app=cloudforms
  3. Navigate to the reported URL/host on a web browser (in this example, cfme.apps.e2e.example.com).
  4. Enter the default CloudForms credentials (Username: admin | Password: smartvm) for the initial login.
  5. Click Login.

Chapter 2. Managing Red Hat CloudForms with OpenShift

This section includes common tasks to manage your Red Hat CloudForms deployment from OpenShift.

2.1. Configuring Custom SSL Certificates for CloudForms

By default, the route that is deployed as part of the template uses edge termination and the certificates that OpenShift is installed with. It is possible to change this in the OpenShift UI with the following steps:

  1. Navigate to ApplicationsRoutes.
  2. Click on the route named httpd, then select ActionsEdit.
  3. Scroll down to the Certificates section. Here you can upload or paste the required certificate files.
  4. Click Save.

2.2. Scaling CloudForms Appliances

StatefulSets in OpenShift manage the deployment and scaling of a set of pods (in this case, CloudForms appliances). StatefulSets ensure ordering that applications will come up by providing unique identities for pods.

Important

Each new replica (server) consumes a physical volume. Before scaling, ensure you have enough physical volumes available to scale.

The following example shows scaling using StatefulSets:

Example: Scaling to two replicas

$ oc scale statefulset cloudforms --replicas=2
statefulset "cloudforms" scaled
$ oc get pods
NAME                 READY     STATUS    RESTARTS   AGE
cloudforms-0           1/1       Running   0          34m
cloudforms-1           1/1       Running   0          5m
memcached-1-mzeer    1/1       Running   0          1h
postgresql-1-dufgp   1/1       Running   0          1h

The newly created replicas will join the existing CloudForms region. Each new pod is numbered in the order it is deployed, starting with 0 and increasing sequentially. For example, replicas in a StatefulSet will be numbered cloudforms-0 cloudforms-1, and so on.

2.3. Creating a Backup

Create a persistent volume for backups using the PV backup template (cfme-pv-backup-example.yaml) in case you need to restore to a previous version.

  1. Create the persistent volume for the backup:

    $ oc create -f cfme-pv-backup-example.yaml
  2. Create the backup persistent volume claim (PVC):

    $ oc create -f cfme-backup-pvc.yaml
  3. Verify the persistent volume claim was created:

    $ oc get pvc
  4. Back up secrets, such as database encryption keys and credentials.

    Important

    Be careful to back up secrets in a secure location.

    $ oc get secret -o yaml --export=true > secrets.yaml
    $ oc get pvc -o yaml --export=true > pvc.yaml
  5. Initiate the database backup:

    $ oc create -f cfme-backup-job.yaml

This step creates a container, and connects to the database pod, pg_basebackup.

2.4. Restoring to a Backup

You can restore to a database backup created in Section 2.3, “Creating a Backup” using the restore template, cfme-restore-job.yaml.

The restore job will look for cfme-backup and cfme-postgresql PVs by default, and the latest successful backup will be restored by default. If existing data is found on the cfme-postgresql volume, it will be renamed and left on the volume.

Important

You must perform a database restore on an offline environment. All pods must be scaled down to 0, and not running.

  1. Shut down all pods:

    $ oc stop all pods
  2. To initiate the database restore, create the restore template:

    $ oc create -f cfme-restore-job.yaml

After the restore job is complete, you can scale the pods back up.

2.5. Uninstalling Red Hat CloudForms from a Project

If no longer needed, you can uninstall the Red Hat CloudForms pod from your project. Note the following commands do not remove SCC permissions, or the project itself.

Important

Use this procedure if only Red Hat CloudForms exists in the project.

  1. Inside the project, run the following as a regular user:

    $ oc delete all --all
  2. Wait approximately 30 seconds for the command to process, then run:

    $ oc delete pvc --all

Chapter 3. Troubleshooting Deployment

Under normal circumstances, the deployment process takes approximately 10 minutes. If the deployment is unsuccessful, examining deployment events and pod logs can help identify any issues.

  1. As a regular user, first retry the failed deployment:

    $ oc get pods
    NAME                 READY     STATUS    RESTARTS   AGE
    cloudforms-1-deploy  0/1       Error     0          25m
    memcached-1-yasfq    1/1       Running   0          24m
    postgresql-1-wfv59   1/1       Running   0          24m
    $ oc deploy cloudforms --retry
    Retried #1
    Use 'oc logs -f dc/cloudforms' to track its progress.
  2. Allow a few seconds for the failed pod to get re-scheduled, then check events and logs:

    $ oc describe pods <pod-name>
    ...
    Events:
      FirstSeen	LastSeen	Count	From							SubobjectPath			Type		Reason		Message
      ---------	--------	-----	----							-------------			--------	------		-------
    15m		15m		1	{kubelet ocp-eval-node-2.e2e.example.com}	spec.containers{cloudforms}	Warning		Unhealthy	Readiness probe failed: Get http://10.1.1.5:80/: dial tcp 10.1.1.5:80: getsockopt: connection refused

    Liveness and readiness probe failures, like in the output above, indicate the pod is taking longer than expected to come online. In this case, check the pod logs.

  3. As the cfme-app container is systemd based, use oc rsh instead of oc logs to obtain journal dumps:

    $ oc rsh <pod-name> journalctl -x
  4. Transferring all logs from the cfme-app pod to a directory on the host for further examination can be useful for troubleshooting. Transfer the logs with the oc rsync command:

    $ oc rsync <pod-name>:/persistent/container-deploy/log \
    /tmp/fail-logs/
    receiving incremental file list
    log/
    log/appliance_initialize_1477272109.log
    log/restore_pv_data_1477272010.log
    log/sync_pv_data_1477272010.log
    
    sent 72 bytes  received 1881 bytes  1302.00 bytes/sec
    total size is 1585  speedup is 0.81