Chapter 20. Speeding up an overcloud upgrade
To speed up the overcloud upgrade process, you can upgrade the control plane incrementally, or upgrade all of your nodes at once.
Upgrade incrementally
You can upgrade 1/3 of your control plane at a time. After you upgrade the first 1/3 of the control plane, you can move your environment into mixed-mode where the control plane APIs are running and the cloud is operational. High availability operational performance can be resumed only after the entire control plane has been upgraded.
Sections 20.1 to 20.4 contain an example upgrade process for an overcloud environment that includes the following node types with composable roles:
- Three Controller nodes
- Three Database nodes
- Three Networker nodes
- Three Ceph Storage nodes
- Multiple Compute nodes
Upgrade the entire overcloud at once
By upgrading the whole overcloud at once, you complete the upgrade much faster than upgrading incrementally. Note that this option requires you to take your control plane and data plane offline.
To upgrade the entire overcloud, see Section 20.5, “Upgrading the entire overcloud at once”.
20.1. Running the overcloud upgrade preparation
The upgrade requires running openstack overcloud upgrade prepare
command, which performs the following tasks:
- Updates the overcloud plan to OpenStack Platform 16.2
- Prepares the nodes for the upgrade
If you are not using the default stack name (overcloud
), set your stack name with the --stack STACK NAME
option replacing STACK NAME
with the name of your stack.
Procedure
Source the
stackrc
file:$ source ~/stackrc
Run the upgrade preparation command:
$ openstack overcloud upgrade prepare \ --stack STACK NAME \ --templates \ -e ENVIRONMENT FILE … -e /home/stack/templates/upgrades-environment.yaml \ -e /home/stack/templates/rhsm.yaml \ -e /home/stack/containers-prepare-parameter.yaml \ -e /usr/share/openstack-tripleo-heat-templates/environments/services/neutron-ovs.yaml \ …
Include the following options relevant to your environment:
-
The environment file (
upgrades-environment.yaml
) with the upgrade-specific parameters (-e
). -
The environment file (
rhsm.yaml
) with the registration and subscription parameters (-e
). -
The environment file (
containers-prepare-parameter.yaml
) with your new container image locations (-e
). In most cases, this is the same environment file that the undercloud uses. -
The environment file (
neutron-ovs.yaml
) to maintain OVS compatibility. -
Any custom configuration environment files (
-e
) relevant to your deployment. -
If applicable, your custom roles (
roles_data
) file using--roles-file
. -
If applicable, your composable network (
network_data
) file using--networks-file
. -
If you use a custom stack name, pass the name with the
--stack
option.
-
The environment file (
- Wait until the upgrade preparation completes.
Download the container images:
$ openstack overcloud external-upgrade run --stack STACK NAME --tags container_image_prepare
20.2. Upgrading the control plane nodes
To upgrade the control plane nodes in your environment to OpenStack Platform 16.2, you must upgrade 1/3 of your control plane nodes at a time, starting with the bootstrap nodes.
During the bootstrap Controller node upgrade process, a new Pacemaker cluster is created and new Red Hat OpenStack 16.2 containers are started on the node, while the remaining Controller nodes continue to run on Red Hat OpenStack 13.
This example includes the following node types with composable roles. The control plane nodes are named using the default overcloud-ROLE-NODEID
convention:
-
overcloud-controller-0
-
overcloud-controller-1
-
overcloud-controller-2
-
overcloud-database-0
-
overcloud-database-1
-
overcloud-database-2
-
overcloud-networker-0
-
overcloud-networker-1
-
overcloud-networker-2
-
overcloud-ceph-0
-
overcloud-ceph-1
-
overcloud-ceph-2
Replace these values for your own node names where applicable.
After you upgrade the overcloud-controller-0
, overcloud-database-0
, overcloud-networker-0
, and overcloud-ceph-0
bootstrap nodes, which comprise the first 1/3 of your control plane nodes, you must upgrade each additional 1/3 of the nodes with Pacemaker services and ensure that each node joins the new Pacemaker cluster started with the bootstrap node. Therefore, you must upgrade overcloud-controller-1
, overcloud-database-1
, overcloud-networker-1
, and overcloud-ceph-1
before you upgrade overcloud-controller-2
, overcloud-database-2
, overcloud-networker-2
, and overcloud-ceph-2
.
Procedure
-
Log in to the undercloud host as the
stack
user. Source the
stackrc
file:$ source ~/stackrc
On the undercloud node, identify the bootstrap Controller node:
$ tripleo-ansible-inventory --list [--stack <stack_name>] |jq .overcloud_Controller.hosts[0]
-
Optional: Replace
<stack_name>
with the name of the stack. If not specified, the default isovercloud
.
-
Optional: Replace
Upgrade the
overcloud-controller-0
,overcloud-database-0
,overcloud-networker-0
, andovercloud-ceph-0
control plane nodes:Run the external upgrade command with the
ceph_systemd
tag:$ openstack overcloud external-upgrade run --stack <stack_name> --tags ceph_systemd -e ceph_ansible_limit=overcloud-controller-0,overcloud-database-0,overcloud-networker-0,overcloud-ceph-0
This command performs the following actions:
- Changes the systemd units that control the Ceph Storage containers to use Podman management.
-
Limits actions to the selected nodes using the
ceph_ansible_limit
variable.
This step is a preliminary measure to prepare the Ceph Storage services for the Leapp upgrade.
Perform a Leapp upgrade of the operating system on each control plane node:
$ openstack overcloud upgrade run --stack <stack_name> --tags system_upgrade --limit overcloud-controller-0,overcloud-database-0,overcloud-networker-0,overcloud-ceph-0
The control plane nodes are rebooted as a part of the Leapp upgrade.
ImportantIf the Ceph node upgrade fails, ensure that
controller-0
has finished upgrading before you proceed with the rest of the upgrade.Copy the latest version of the database from an existing node to the bootstrap node:
$ openstack overcloud external-upgrade run --stack <stack_name> --tags system_upgrade_transfer_data
ImportantThis command causes an outage on the control plane. You cannot perform any standard operations on the overcloud during the next few steps.
Run the upgrade command with the
nova_hybrid_state
tag and run only theupgrade_steps_playbook.yaml
playbook:$ openstack overcloud upgrade run --stack <stack_name> \ --playbook upgrade_steps_playbook.yaml \ --tags nova_hybrid_state --limit all
This command launches temporary 16.2 containers on Compute nodes to help facilitate workload migration when you upgrade Compute nodes at a later step.
Run the upgrade command with no tags:
$ openstack overcloud upgrade run --stack <stack_name> --limit overcloud-controller-0,overcloud-database-0,overcloud-networker-0,overcloud-ceph-0 --playbook all
This command performs the Red Hat OpenStack Platform upgrade.
ImportantThe control plane becomes active when this command finishes. You can perform standard operations on the overcloud again.
Optional: On the bootstrap Controller node, verify that after the upgrade, the new Pacemaker cluster is started and that the control plane services such as galera, rabbit, haproxy, and redis are running:
$ sudo pcs status
Upgrade the
overcloud-controller-1
,overcloud-database-1
,overcloud-networker-1
, andovercloud-ceph-1
control plane nodes:Log in to the
overcloud-controller-1
node and verify that the old cluster is no longer running:$ sudo pcs status
An error similar to the following is displayed when the cluster is not running:
Error: cluster is not currently running on this node
Run the external upgrade command with the
ceph_systemd
tag:$ openstack overcloud external-upgrade run --stack <stack_name> --tags ceph_systemd -e ceph_ansible_limit=overcloud-controller-1,overcloud-database-1,overcloud-networker-1,overcloud-ceph-1
This command performs the following functions:
- Changes the systemd units that control the Ceph Storage containers to use Podman management.
-
Limits actions to the selected nodes using the
ceph_ansible_limit
variable.
This step is a preliminary measure to prepare the Ceph Storage services for the Leapp upgrade.
Run the upgrade command with the
system_upgrade
tag:$ openstack overcloud upgrade run --stack <stack_name> --tags system_upgrade --limit overcloud-controller-1,overcloud-database-1,overcloud-networker-1,overcloud-ceph-1
This command performs the following actions:
- Performs a Leapp upgrade of the operating system.
- Performs a reboot as a part of the Leapp upgrade.
Run the upgrade command with no tags:
$ openstack overcloud upgrade run --stack <stack_name> --limit overcloud-controller-0,overcloud-controller-1,overcloud-database-0,overcloud-database-1,overcloud-networker-0,overcloud-networker-1,overcloud-ceph-0,overcloud-ceph-1
This command performs the Red Hat OpenStack Platform upgrade. In addition to this node, include the previously upgraded bootstrap nodes in the
--limit
option.
Upgrade the
overcloud-controller-2
,overcloud-database-2
,overcloud-networker-2
, andovercloud-ceph-2
control plane nodes:Log in to the
overcloud-controller-2
node and verify that the old cluster is no longer running:$ sudo pcs status
An error similar to the following is displayed when the cluster is not running:
Error: cluster is not currently running on this node
Run the external upgrade command with the
ceph_systemd
tag:$ openstack overcloud external-upgrade run --stack <stack_name> --tags ceph_systemd -e ceph_ansible_limit=overcloud-controller-2,overcloud-database-2,overcloud-networker-2,overcloud-ceph-2
This command performs the following functions:
- Changes the systemd units that control the Ceph Storage containers to use Podman management.
-
Limits actions to the selected nodes using the
ceph_ansible_limit
variable.
This step is a preliminary measure to prepare the Ceph Storage services for the Leapp upgrade.
Run the upgrade command with the
system_upgrade
tag:$ openstack overcloud upgrade run --stack <stack_name> --tags system_upgrade --limit overcloud-controller-2,overcloud-database-2,overcloud-networker-2,overcloud-ceph-2
This command performs the following actions:
- Performs a Leapp upgrade of the operating system.
- Performs a reboot as a part of the Leapp upgrade.
Run the upgrade command with no tags:
$ openstack overcloud upgrade run --stack <stack_name> --limit overcloud-controller-0,overcloud-controller-1,overcloud-controller-2,overcloud-database-0,overcloud-database-1,overcloud-database-2,overcloud-networker-0,overcloud-networker-1,overcloud-networker-2,overcloud-ceph-0,overcloud-ceph-1,overcloud-ceph-2
This command performs the Red Hat OpenStack Platform upgrade. Include all control plane nodes in the
--limit
option.
20.3. Upgrading Compute nodes in parallel
To upgrade a large number of Compute nodes to OpenStack Platform 16.2, you can run the openstack overcloud upgrade run
command with the --limit Compute
option in parallel on groups of 20 nodes.
You can run multiple upgrade tasks in the background, where each task upgrades a separate group of 20 nodes. When you use this method to upgrade Compute nodes in parallel, you cannot select which nodes you upgrade. The selection of nodes is based on the inventory file that you generate when you run the tripleo-ansible-inventory
command. For example, if you have 80 Compute nodes in your deployment, you can run the following commands to update the Compute nodes in parallel:
$ openstack overcloud upgrade run -y --limit 'Compute[0:19]' > upgrade-compute-00-19.log 2>&1 & $ openstack overcloud upgrade run -y --limit 'Compute[20:29]' > upgrade-compute-20-29.log 2>&1 & $ openstack overcloud upgrade run -y --limit 'Compute[40:59]' > update-compute-40-59.log 2>&1 & $ openstack overcloud upgrade run -y --limit 'Compute[60:79]' > update-compute-60-79.log 2>&1 &
To upgrade specific Compute nodes, use a comma-separated list of nodes:
$ openstack overcloud upgrade run --limit <Compute0>,<Compute1>,<Compute2>,<Compute3>
If you are not using the default stack name overcloud
, use the --stack STACK NAME
option and replace STACK NAME
with name of your stack.
Procedure
-
Log in to the undercloud host as the
stack
user. Source the
stackrc
file:$ source ~/stackrc
- Migrate your instances. For more information on migration strategies, see Migrating virtual machines between Compute nodes.
Run the upgrade command with the
system_upgrade
tag:$ openstack overcloud upgrade run -y --stack STACK NAME --tags system_upgrade --limit 'Compute[0:19]' > upgrade-compute-00-19.log 2>&1 & $ openstack overcloud upgrade run -y --stack STACK NAME --tags system_upgrade --limit 'Compute[20:29]' > upgrade-compute-20-29.log 2>&1 & $ openstack overcloud upgrade run -y --stack STACK NAME --tags system_upgrade --limit 'Compute[40:59]' > update-compute-40-59.log 2>&1 & $ openstack overcloud upgrade run -y --stack STACK NAME --tags system_upgrade --limit 'Compute[60:79]' > update-compute-60-79.log 2>&1 &
This command performs the following actions:
- Performs a Leapp upgrade of the operating system.
- Performs a reboot as a part of the Leapp upgrade.
Run the upgrade command with no tags:
$ openstack overcloud upgrade run -y --stack STACK NAME --limit 'Compute[0:19]' > upgrade-compute-00-19.log 2>&1 & $ openstack overcloud upgrade run -y --stack STACK NAME --limit 'Compute[20:29]' > upgrade-compute-20-29.log 2>&1 & $ openstack overcloud upgrade run -y --stack STACK NAME --limit 'Compute[40:59]' > update-compute-40-59.log 2>&1 & $ openstack overcloud upgrade run -y --stack STACK NAME --limit 'Compute[60:79]' > update-compute-60-79.log 2>&1 &
This command performs the Red Hat OpenStack Platform upgrade.
Optional: To upgrade selected Compute nodes, use the
--limit
option with a comma-separated list of nodes that you want to upgrade. The following example upgrades theovercloud-compute-0
,overcloud-compute-1
,overcloud-compute-2
nodes in parallel.Run the upgrade command with the
system_upgrade
tag:$ openstack overcloud upgrade run --stack STACK NAME --tags system_upgrade --limit overcloud-compute-0,overcloud-compute-1,overcloud-compute-2
Run the upgrade command with no tags:
$ openstack overcloud upgrade run --stack STACK NAME --limit overcloud-compute-0,overcloud-compute-1,overcloud-compute-2
20.4. Synchronizing the overcloud stack
The upgrade requires an update the overcloud stack to ensure that the stack resource structure and parameters align with a fresh deployment of OpenStack Platform 16.2.
If you are not using the default stack name (overcloud
), set your stack name with the --stack STACK NAME
option replacing STACK NAME
with the name of your stack.
Procedure
Source the
stackrc
file:$ source ~/stackrc
Edit the
containers-prepare-parameter.yaml
file and remove the following parameters and their values:-
ceph3_namespace
-
ceph3_tag
-
ceph3_image
-
name_prefix_stein
-
name_suffix_stein
-
namespace_stein
-
tag_stein
-
-
To re-enable fencing in your overcloud, set the
EnableFencing
parameter totrue
in thefencing.yaml
environment file. Run the upgrade finalization command:
$ openstack overcloud upgrade converge \ --stack STACK NAME \ --templates \ -e ENVIRONMENT FILE … -e /home/stack/templates/upgrades-environment.yaml \ -e /home/stack/templates/rhsm.yaml \ -e /home/stack/containers-prepare-parameter.yaml \ -e /usr/share/openstack-tripleo-heat-templates/environments/services/neutron-ovs.yaml \ …
Include the following options relevant to your environment:
-
The environment file (
upgrades-environment.yaml
) with the upgrade-specific parameters (-e
). -
The environment file (
fencing.yaml
) with theEnableFencing
parameter set totrue
. -
The environment file (
rhsm.yaml
) with the registration and subscription parameters (-e
). -
The environment file (
containers-prepare-parameter.yaml
) with your new container image locations (-e
). In most cases, this is the same environment file that the undercloud uses. -
The environment file (
neutron-ovs.yaml
) to maintain OVS compatibility. -
Any custom configuration environment files (
-e
) relevant to your deployment. -
If applicable, your custom roles (
roles_data
) file using--roles-file
. -
If applicable, your composable network (
network_data
) file using--networks-file
. -
If you use a custom stack name, pass the name with the
--stack
option.
-
The environment file (
- Wait until the stack synchronization completes.
You do not need the upgrades-environment.yaml
file for any further deployment operations.
20.5. Upgrading the entire overcloud at once
This upgrade process requires you to first shut down all workloads that are running in the overcloud. You then upgrade all of your overcloud systems, and restart the workloads afterward. This process is driven from the undercloud.
You can also upgrade Compute nodes that include Red Hat Ceph Storage as part of this procedure, or upgrade them separately after you upgrade all the other nodes.
Prerequisites
In the
upgrades-environment.yaml
file, include the following parameter inparameter_defaults
:AllInOneUpgrade: true
Procedure
- Shut down your workloads.
If you deployed director-integrated Ceph, switch the Ceph systemd files to podman:
$ openstack overcloud external-upgrade run --stack overcloud --tags ceph_systemd -e ceph_ansible_limit=controller-0,controller-1,controller-2,ceph-0,ceph-1,ceph-2,networker-0,networker-1
-
Replace
controller-0
,controller-1
,controller-2
,ceph-0
,ceph-1
,ceph-2
,networker-0
,networker-1
with the server names for the roles in your environment. To upgrade Compute nodes that contain Ceph, include the hostname of the Compute node in the
openstack overcloud external-upgrade run
command. For example:$ openstack overcloud upgrade run --stack overcloud --tags ceph_systemd -e ceph_ansible_limit=overcloud-compute-0,overcloud-compute-1
Additionally, include the hostname of the Compute node in the commands in steps 4 and 5.
-
Replace
Stop all RHOSP services on the nodes:
$ openstack overcloud external-upgrade run --stack overcloud --tags system_upgrade_stop_services
Run the system upgrade on all nodes. If you deployed director-integrated Ceph, include all Ceph nodes in the same --limit command:
$ openstack overcloud upgrade run --stack overcloud --tags system_upgrade --limit controller-0,controller-1,controller-2,ceph-0,ceph-1,ceph-2,networker-0,networker-1
Run the upgrade without tags:
$ openstack overcloud upgrade run --stack overcloud --limit controller-0,controller-1,controller-2,ceph-0,ceph-1,ceph-2,networker-0,networker-1
Next steps
- Continue with the upgrade from Synchronizing the overcloud stack.