Chapter 1. Introduction to upgrading OpenStack Platform 15 to 16.1

This document provides a workflow to help you upgrade your Red Hat OpenStack Platform 15 to 16.1 and update your environment with the latest packages and containers.

This guide provides an upgrade path through the following versions:

Old OpenStack VersionNew OpenStack Version

Red Hat OpenStack Platform 15.0

Red Hat OpenStack Platform 16.1

1.1. High level workflow

The following table provides an outline of the tasks required for the upgrade process:

TaskDescription

Preparing your environment for the upgrade

Prepare and configure the repositories, modules, and containers that are required for the upgrade.

Upgrading the undercloud

Upgrade the undercloud to the latest OpenStack Platform 16.1 version.

Upgrading the overcloud

Upgrade the overcloud to the latest OpenStack Platform 16.1 version. Upgrade all Controller, Compute, and Ceph Storage services. Run the convergence command to refresh your overcloud stack.

Rebooting the overcloud

Reboot the overcloud.

1.2. Known issues that might block an upgrade

Review the following list of known issues that might affect the upgrade from OpenStack Platform 15 to 16.1.

BZ#1895220 - Network communication problems for instances using OVN provider network
Updates to 16.1.3 from an earlier version (16.0.x or 16.1.x) can cause network disruption due to a database issue with Open Virtual Network (OVN). When you perform a minor version update, you normally update Controller nodes before Compute nodes. When you update Controller nodes, director updates the ovndb-north database schema to the latest version. The ovn-controller service on Compute nodes cannot interpret the newer version of the ovndb-north database schema and cannot obtain the correct network flow for instances. As a workaround to minimize the network disruption, you must update the ovn_controller service on Compute nodes before you run the openstack overcloud update run command and after you run the openstack overcloud update prepare command. For more information, see the "OVN update in 16.1 workaround" knowledgebase article. Red Hat aims to resolve this issue in the next 16.1 minor release update.
BZ#1872404 - restarting nodes in parallel while maintaining quorum creates an unexpected node shutdown
Until this issue is resolved, for nodes based on composable roles, you must update the Database role first, before you can update Controller, Messaging, Compute, Ceph, and other roles.