Increase disk size of controllers to accommodate large images in new deployments

Products

• VMware Integrated OpenStack

Issue/Introduction

When uploading images from the Horizon UI, the images will get stored in the glance-api pods on those controllers.  If the image is larger than the remaining disk size the operation will fail.

The default size for the controller nodes that run the glance-api pods are:

 sizing:
      small:
        numCPUs:    4
        memoryMB:   16384
        diskSizeGB: 25
      medium:
        numCPUs:    8
        memoryMB:   32768
        diskSizeGB: 50
      large:
        numCPUs:    12
        memoryMB:   32768
        diskSizeGB: 75

To accommodate larger images, we need to change the disk size of the controller nodes.

Environment

6.x

Resolution

For new VMware Integrated Openstack Deployments

1. Install the VMware Integrated OpenStack Virtual Appliance
2. Once the appliance has been successfully deployed access the OpenStack Management Server via SSH.
3. The controller resource configuration is maintained in a K8s configmap resource named lcm-config. Open the lcm-config configmap for editing in vi using the following command:

kubectl -n openstack edit configmap lcm-config

4. In the configmap locate the controller resource size. The diskSizeGB value is used to determine how much disk space a controller will be allocated. Change the diskSizeGB value to the desired disk space in GB.

data:
  lcm.yaml: |
    sizing:
      small:
        numCPUs: 4
        memoryMB: 16384
        diskSizeGB: 25
      medium:
        numCPUs: 8
        memoryMB: 32768
        diskSizeGB: 50
      large:
        numCPUs: 12
        memoryMB: 32768
        diskSizeGB: 75

5. Save the file.
6. Create an OpenStack Deployment. 

Note:  Make sure to select the sizing you changed in step 4 in the wizard.

Additional Information

Extending root disk of VIO controllers