Creating the initial volume group as part of the deployment by using a file-backed loop device

The initial cinder-volumes volume group can be created as part of deploying your topology. Use these instructions to create the initial volume group that uses a file-backed loop device.

About this task

When you create a volume group that is backed by a file on the block storage node, the deployment process creates a file, /var/lib/cinder/cinder-volumes.img, and attaches a loop device, /dev/loopx, to the file. Additionally, it creates a service, cinder-group-active, that is used to reattach the loop device when the system is rebooted.

The following information describes details about supported options.
Table 1. Summary of support details
Support options Details
Supported hypervisor types and architectures? KVM or QEMU on x86_64, PowerKVM, and Hyper-V
Support for post-deployment customization? Yes
Supported topologies? All, except HA controller +n compute
Storage (Cinder) node? Controller or block storage node
Operating system and architecture on storage node? Red Hat Enterprise Linux® 7.1, 7.2, or 7.3 on x86_64 or ppc64

Procedure

  1. 1. Modify the environment or node attributes file to contain the following additional attributes. These attributes are in the override_attributes section of the environment file for the Minimal topology and in the default_attributes section for other topologies.
    • openstack.block-storage.volume.create_volume_group: Set this value to true.
    • openstack.block-storage.volume.create_volume_group_type : Set this value to "file".
    • openstack.block-storage.volume.volume_group_size : Set this value to the size of the volume group in GB. There must be enough free space on the file system to create the file. The default value is 40.
  2. When complete, return to the relevant topology deployment or update process and complete the remaining steps.