Migrating KVM compute nodes

1. To upgrade a compute node from Red Hat Enterprise Linux 6.5 to Red Hat Enterprise Linux 7.1, 7.2, or 7.3 migrate all the instances from the source compute node that you plan to upgrade to another compute node. Ensure that the target compute node has enough resource to accept all the instances from the source node. You can migrate instances from one Red Hat Enterprise Linux 6.5 compute node to another Red Hat Enterprise Linux 6.5 compute node to save the system resource, and then upgrade the source node to Red Hat Enterprise Linux 7.1, 7.2, or 7.3. Alternatively, you can deploy a new compute node with Red Hat Enterprise Linux 7.1, 7.2, or 7.3 at the beginning of the migration to function as the target node.
   Note: If you want the new Red Hat Enterprise Linux 7.1, 7.2, or 7.3 compute nodes to have the same environment as the controller, you must disable Consistent Network Device Naming on the compute nodes in the same manner as you disabled it for the controller node when you completed the instructions for Deploying a new controller. For the compute nodes, you can disable Consistent Network Device Naming by copying the controller environment and renaming the environment for compute node usage. Alternatively, you can use an attribute file to override network information of node-attribute. For more information about attribute files, see Deploying an advanced configuration with KVM or QEMU compute nodes
2. After all the instances are migrated off the source node to other compute nodes, you can reinstall the source node to Red Hat Enterprise Linux 7.1, 7.2, or 7.3 and redeploy this node as a new compute node. Note: This step is not required if you have enough system resources to do the migration directly to a new Red Hat Enterprise Linux 7.1, 7.2, or 7.3 compute node.

Before you begin, ensure that you completed the Migrating the OpenStack controller steps.