$ sudo openstack quota set --secgroups 250 --secgroup-rules 1000 --ports 1500 --subnets 250 --networks 250 <project>
install-config.yaml
file for OpenStack with KuryrIn OKD version 4.9, you can install a cluster on OpenStack that runs on user-provisioned infrastructure.
Using your own infrastructure allows you to integrate your cluster with existing infrastructure and modifications. The process requires more labor on your part than installer-provisioned installations, because you must create all OpenStack resources, like Nova servers, Neutron ports, and security groups. However, Red Hat provides Ansible playbooks to help you in the deployment process.
You reviewed details about the OKD installation and update processes.
You read the documentation on selecting a cluster installation method and preparing it for users.
You verified that OKD 4.9 is compatible with your OpenStack version by using the Supported platforms for OpenShift clusters section. You can also compare platform support across different versions by viewing the OKD on OpenStack support matrix.
You have an OpenStack account where you want to install OKD.
On the machine from which you run the installation program, you have:
A single directory in which you can keep the files you create during the installation process
Python 3
Kuryr is a container network interface (CNI) plugin solution that uses the Neutron and Octavia OpenStack services to provide networking for pods and Services.
Kuryr and OKD integration is primarily designed for OKD clusters running on OpenStack VMs. Kuryr improves the network performance by plugging OKD pods into OpenStack SDN. In addition, it provides interconnectivity between pods and OpenStack virtual instances.
Kuryr components are installed as pods in OKD using the
openshift-kuryr
namespace:
kuryr-controller
- a single service instance installed on a master
node.
This is modeled in OKD as a Deployment
object.
kuryr-cni
- a container installing and configuring Kuryr as a CNI driver on
each OKD node. This is modeled in OKD as a DaemonSet
object.
The Kuryr controller watches the OKD API server for pod, service, and namespace create, update, and delete events. It maps the OKD API calls to corresponding objects in Neutron and Octavia. This means that every network solution that implements the Neutron trunk port functionality can be used to back OKD via Kuryr. This includes open source solutions such as Open vSwitch (OVS) and Open Virtual Network (OVN) as well as Neutron-compatible commercial SDNs.
Kuryr is recommended for OKD deployments on encapsulated OpenStack tenant networks to avoid double encapsulation, such as running an encapsulated OKD SDN over an OpenStack network.
If you use provider networks or tenant VLANs, you do not need to use Kuryr to avoid double encapsulation. The performance benefit is negligible. Depending on your configuration, though, using Kuryr to avoid having two overlays might still be beneficial.
Kuryr is not recommended in deployments where all of the following criteria are true:
The OpenStack version is less than 16.
The deployment uses UDP services, or a large number of TCP services on few hypervisors.
or
The ovn-octavia
Octavia driver is disabled.
The deployment uses a large number of TCP services on few hypervisors.
When using Kuryr SDN, the pods, services, namespaces, and network policies are using resources from the OpenStack quota; this increases the minimum requirements. Kuryr also has some additional requirements on top of what a default install requires.
Use the following quota to satisfy a default cluster’s minimum requirements:
Resource | Value |
---|---|
Floating IP addresses |
3 - plus the expected number of Services of LoadBalancer type |
Ports |
1500 - 1 needed per Pod |
Routers |
1 |
Subnets |
250 - 1 needed per Namespace/Project |
Networks |
250 - 1 needed per Namespace/Project |
RAM |
112 GB |
vCPUs |
28 |
Volume storage |
275 GB |
Instances |
7 |
Security groups |
250 - 1 needed per Service and per NetworkPolicy |
Security group rules |
1000 |
Load balancers |
100 - 1 needed per Service |
Load balancer listeners |
500 - 1 needed per Service-exposed port |
Load balancer pools |
500 - 1 needed per Service-exposed port |
A cluster might function with fewer than recommended resources, but its performance is not guaranteed.
If OpenStack object storage (Swift) is available and operated by a user account with the |
If you are using OpenStack version 16 with the Amphora driver rather than the OVN Octavia driver, security groups are associated with service accounts instead of user projects. |
Take the following notes into consideration when setting resources:
The number of ports that are required is larger than the number of pods. Kuryr uses ports pools to have pre-created ports ready to be used by pods and speed up the pods' booting time.
Each network policy is mapped into an OpenStack security group, and
depending on the NetworkPolicy
spec, one or more rules are added to the
security group.
Each service is mapped to an OpenStack load balancer. Consider this requirement when estimating the number of security groups required for the quota.
If you are using
OpenStack version 15 or earlier, or the ovn-octavia driver
, each load balancer
has a security group with the user project.
The quota does not account for load balancer resources (such as VM resources), but you must consider these resources when you decide the OpenStack deployment’s size. The default installation will have more than 50 load balancers; the clusters must be able to accommodate them.
If you are using OpenStack version 16 with the OVN Octavia driver enabled, only one load balancer VM is generated; services are load balanced through OVN flows.
An OKD deployment comprises control plane machines, compute machines, and a bootstrap machine.
To enable Kuryr SDN, your environment must meet the following requirements:
Run OpenStack 13+.
Have Overcloud with Octavia.
Use Neutron Trunk ports extension.
Use openvswitch
firewall driver if ML2/OVS Neutron driver is used instead
of ovs-hybrid
.
When using Kuryr SDN, you must increase quotas to satisfy the OpenStack resources used by pods, services, namespaces, and network policies.
Increase the quotas for a project by running the following command:
$ sudo openstack quota set --secgroups 250 --secgroup-rules 1000 --ports 1500 --subnets 250 --networks 250 <project>
Kuryr CNI leverages the Neutron Trunks extension to plug containers into the
OpenStack SDN, so you must use the trunks
extension for Kuryr to properly work.
In addition, if you leverage the default ML2/OVS Neutron driver, the firewall
must be set to openvswitch
instead of ovs_hybrid
so that security groups are
enforced on trunk subports and Kuryr can properly handle network policies.
Kuryr SDN uses OpenStack’s Octavia LBaaS to implement OKD services. Thus, you must install and configure Octavia components in OpenStack to use Kuryr SDN.
To enable Octavia, you must include the Octavia service during the installation of the OpenStack Overcloud, or upgrade the Octavia service if the Overcloud already exists. The following steps for enabling Octavia apply to both a clean install of the Overcloud or an Overcloud update.
The following steps only capture the key pieces required during the deployment of OpenStack when dealing with Octavia. It is also important to note that registry methods vary. This example uses the local registry method. |
If you are using the local registry, create a template to upload the images to the registry. For example:
(undercloud) $ openstack overcloud container image prepare \
-e /usr/share/openstack-tripleo-heat-templates/environments/services-docker/octavia.yaml \
--namespace=registry.access.redhat.com/rhosp13 \
--push-destination=<local-ip-from-undercloud.conf>:8787 \
--prefix=openstack- \
--tag-from-label {version}-{product-version} \
--output-env-file=/home/stack/templates/overcloud_images.yaml \
--output-images-file /home/stack/local_registry_images.yaml
Verify that the local_registry_images.yaml
file contains the Octavia images.
For example:
...
- imagename: registry.access.redhat.com/rhosp13/openstack-octavia-api:13.0-43
push_destination: <local-ip-from-undercloud.conf>:8787
- imagename: registry.access.redhat.com/rhosp13/openstack-octavia-health-manager:13.0-45
push_destination: <local-ip-from-undercloud.conf>:8787
- imagename: registry.access.redhat.com/rhosp13/openstack-octavia-housekeeping:13.0-45
push_destination: <local-ip-from-undercloud.conf>:8787
- imagename: registry.access.redhat.com/rhosp13/openstack-octavia-worker:13.0-44
push_destination: <local-ip-from-undercloud.conf>:8787
The Octavia container versions vary depending upon the specific OpenStack release installed. |
Pull the container images from registry.redhat.io
to the Undercloud node:
(undercloud) $ sudo openstack overcloud container image upload \
--config-file /home/stack/local_registry_images.yaml \
--verbose
This may take some time depending on the speed of your network and Undercloud disk.
Since an Octavia load balancer is used to access the OKD API, you must increase their listeners' default timeouts for the connections. The default timeout is 50 seconds. Increase the timeout to 20 minutes by passing the following file to the Overcloud deploy command:
(undercloud) $ cat octavia_timeouts.yaml
parameter_defaults:
OctaviaTimeoutClientData: 1200000
OctaviaTimeoutMemberData: 1200000
This is not needed for OpenStack 13.0.13+. |
Install or update your Overcloud environment with Octavia:
$ openstack overcloud deploy --templates \
-e /usr/share/openstack-tripleo-heat-templates/environments/services-docker/octavia.yaml \
-e octavia_timeouts.yaml
This command only includes the files associated with Octavia; it varies based on your specific installation of OpenStack. See the OpenStack documentation for further information. For more information on customizing your Octavia installation, see installation of Octavia using Director. |
When leveraging Kuryr SDN, the Overcloud installation requires the Neutron |
In OpenStack versions earlier than 13.0.13, add the project ID
to the octavia.conf
configuration file after you create the project.
To enforce network policies across services, like when traffic goes through the Octavia load balancer, you must ensure Octavia creates the Amphora VM security groups on the user project.
This change ensures that required load balancer security groups belong to that project, and that they can be updated to enforce services isolation.
This task is unnecessary in OpenStack version 13.0.13 or later. Octavia implements a new ACL API that restricts access to the load balancers VIP. |
Get the project ID
$ openstack project show <project>
+-------------+----------------------------------+
| Field | Value |
+-------------+----------------------------------+
| description | |
| domain_id | default |
| enabled | True |
| id | PROJECT_ID |
| is_domain | False |
| name | *<project>* |
| parent_id | default |
| tags | [] |
+-------------+----------------------------------+
Add the project ID to octavia.conf
for the controllers.
Source the stackrc
file:
$ source stackrc # Undercloud credentials
List the Overcloud controllers:
$ openstack server list
+--------------------------------------+--------------+--------+-----------------------+----------------+------------+
│
| ID | Name | Status | Networks
| Image | Flavor |
│
+--------------------------------------+--------------+--------+-----------------------+----------------+------------+
│
| 6bef8e73-2ba5-4860-a0b1-3937f8ca7e01 | controller-0 | ACTIVE |
ctlplane=192.168.24.8 | overcloud-full | controller |
│
| dda3173a-ab26-47f8-a2dc-8473b4a67ab9 | compute-0 | ACTIVE |
ctlplane=192.168.24.6 | overcloud-full | compute |
│
+--------------------------------------+--------------+--------+-----------------------+----------------+------------+
SSH into the controller(s).
$ ssh heat-admin@192.168.24.8
Edit the octavia.conf
file to add the project into the list of projects where
Amphora security groups are on the user’s account.
# List of project IDs that are allowed to have Load balancer security groups # belonging to them. amp_secgroup_allowed_projects = PROJECT_ID
Restart the Octavia worker so the new configuration loads.
controller-0$ sudo docker restart octavia_worker
Depending on your OpenStack environment, Octavia might not support UDP listeners. If you use Kuryr SDN on OpenStack version 13.0.13 or earlier, UDP services are not supported. OpenStack version 16 or later support UDP. |
Octavia supports multiple provider drivers through the Octavia API.
To see all available Octavia provider drivers, on a command line, enter:
$ openstack loadbalancer provider list
+---------+-------------------------------------------------+
| name  | description                   |
+---------+-------------------------------------------------+
| amphora | The Octavia Amphora driver. Â Â Â Â Â Â Â Â Â Â |
| octavia | Deprecated alias of the Octavia Amphora driver. |
| ovn   | Octavia OVN driver.               |
+---------+-------------------------------------------------+
Beginning with OpenStack version 16, the Octavia OVN provider driver (ovn
) is supported on
OKD on OpenStack deployments.
ovn
is an integration driver for the load balancing
that Octavia and OVN provide. It supports basic load balancing capabilities,
and is based on OpenFlow rules. The driver is automatically enabled
in Octavia by Director on deployments that use OVN Neutron ML2.
The Amphora provider driver is the default driver. If ovn
is enabled, however, Kuryr uses it.
If Kuryr uses ovn
instead of Amphora, it offers the following benefits:
Decreased resource requirements. Kuryr does not require a load balancer VM for each service.
Reduced network latency.
Increased service creation speed by using OpenFlow rules instead of a VM for each service.
Distributed load balancing actions across all nodes instead of centralized on Amphora VMs.
Using OKD with Kuryr SDN has several known limitations.
Using OKD with Kuryr SDN has several limitations that apply to all versions and environments:
Service
objects with the NodePort
type are not supported.
Clusters that use the OVN Octavia provider driver support Service
objects for which the .spec.selector
property is unspecified only if the .subsets.addresses
property of the Endpoints
object includes the subnet of the nodes or pods.
If the subnet on which machines are created is not connected to a router, or if the subnet is connected, but the router has no external gateway set, Kuryr cannot create floating IPs for Service
objects with type LoadBalancer
.
Configuring the sessionAffinity=ClientIP
property on Service
objects does not have an effect. Kuryr does not support this setting.
Using OKD with Kuryr SDN has several limitations that depend on the OpenStack version.
OpenStack versions before 16 use the default Octavia load balancer driver (Amphora). This driver requires that one Amphora load balancer VM is deployed per OKD service. Creating too many services can cause you to run out of resources.
Deployments of later versions of OpenStack that have the OVN Octavia driver disabled also use the Amphora driver. They are subject to the same resource concerns as earlier versions of OpenStack.
Octavia OpenStack versions before 13.0.13 do not support UDP listeners. Therefore, OKD UDP services are not supported.
Octavia OpenStack versions before 13.0.13 cannot listen to multiple protocols on the same port. Services that expose the same port to different protocols, like TCP and UDP, are not supported.
Kuryr SDN does not support automatic unidling by a service.
There are limitations when using Kuryr SDN that depend on your deployment environment.
Because of Octavia’s lack of support for the UDP protocol and multiple listeners, if the OpenStack version is earlier than 13.0.13, Kuryr forces pods to use TCP for DNS resolution.
In Go versions 1.12 and earlier, applications that are compiled with CGO support disabled use UDP only. In this case,
the native Go resolver does not recognize the use-vc
option in resolv.conf
, which controls whether TCP is forced for DNS resolution.
As a result, UDP is still used for DNS resolution, which fails.
To ensure that TCP forcing is allowed, compile applications either with the environment variable CGO_ENABLED
set to 1
, i.e. CGO_ENABLED=1
, or ensure that the variable is absent.
In Go versions 1.13 and later, TCP is used automatically if DNS resolution using UDP fails.
musl-based containers, including Alpine-based containers, do not support the |
As a result of the OpenStack upgrade process, the Octavia API might be changed, and upgrades to the Amphora images that are used for load balancers might be required.
You can address API changes on an individual basis.
If the Amphora image is upgraded, the OpenStack operator can handle existing load balancer VMs in two ways:
Upgrade each VM by triggering a load balancer failover.
Leave responsibility for upgrading the VMs to users.
If the operator takes the first option, there might be short downtimes during failovers.
If the operator takes the second option, the existing load balancers will not support upgraded Octavia API features, like UDP listeners. In this case, users must recreate their Services to use these features.
If OKD detects a new Octavia version that supports UDP load balancing, it recreates the DNS service automatically. The service recreation ensures that the service default supports UDP load balancing. The recreation causes the DNS service approximately one minute of downtime. |
By default, the OKD installation process creates three control plane machines.
Each machine requires:
An instance from the OpenStack quota
A port from the OpenStack quota
A flavor with at least 16 GB memory and 4 vCPUs
At least 100 GB storage space from the OpenStack quota
By default, the OKD installation process creates three compute machines.
Each machine requires:
An instance from the OpenStack quota
A port from the OpenStack quota
A flavor with at least 8 GB memory and 2 vCPUs
At least 100 GB storage space from the OpenStack quota
Compute machines host the applications that you run on OKD; aim to run as many as you can. |
During installation, a bootstrap machine is temporarily provisioned to stand up the control plane. After the production control plane is ready, the bootstrap machine is deprovisioned.
The bootstrap machine requires:
An instance from the OpenStack quota
A port from the OpenStack quota
A flavor with at least 16 GB memory and 4 vCPUs
At least 100 GB storage space from the OpenStack quota
The Ansible playbooks that simplify the installation process on user-provisioned infrastructure require several Python modules. On the machine where you will run the installer, add the modules' repositories and then download them.
These instructions assume that you are using Fedora 8. |
Python 3 is installed on your machine.
On a command line, add the repositories:
Register with Red Hat Subscription Manager:
$ sudo subscription-manager register # If not done already
Pull the latest subscription data:
$ sudo subscription-manager attach --pool=$YOUR_POOLID # If not done already
Disable the current repositories:
$ sudo subscription-manager repos --disable=* # If not done already
Add the required repositories:
$ sudo subscription-manager repos \
--enable=rhel-8-for-x86_64-baseos-rpms \
--enable=openstack-16-tools-for-rhel-8-x86_64-rpms \
--enable=ansible-2.9-for-rhel-8-x86_64-rpms \
--enable=rhel-8-for-x86_64-appstream-rpms
Install the modules:
$ sudo yum install python3-openstackclient ansible python3-openstacksdk python3-netaddr
Ensure that the python
command points to python3
:
$ sudo alternatives --set python /usr/bin/python3
Download Ansible playbooks that you can use to install OKD on your own OpenStack infrastructure.
The curl command-line tool is available on your machine.
To download the playbooks to your working directory, run the following script from a command line:
$ xargs -n 1 curl -O <<< '
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/bootstrap.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/common.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/compute-nodes.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/control-plane.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/inventory.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/network.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/security-groups.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/down-bootstrap.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/down-compute-nodes.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/down-control-plane.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/down-load-balancers.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/down-network.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/down-security-groups.yaml
https://raw.githubusercontent.com/openshift/installer/release-4.9/upi/openstack/down-containers.yaml'
The playbooks are downloaded to your machine.
During the installation process, you can modify the playbooks to configure your deployment. Retain all playbooks for the life of your cluster. You must have the playbooks to remove your OKD cluster from OpenStack. |
You must match any edits you make in the |
Before you install OKD, download the installation file on a local computer.
You have a computer that runs Linux or macOS, with 500 MB of local disk space
Download installer from https://github.com/openshift/okd/releases
The installation program creates several files on the computer that you use to install your cluster. You must keep the installation program and the files that the installation program creates after you finish installing the cluster. Both files are required to delete the cluster. |
Deleting the files created by the installation program does not remove your cluster, even if the cluster failed during installation. To remove your cluster, complete the OKD uninstallation procedures for your specific cloud provider. |
Extract the installation program. For example, on a computer that uses a Linux operating system, run the following command:
$ tar -xvf openshift-install-linux.tar.gz
Download your installation pull secret from the Red Hat OpenShift Cluster Manager. This pull secret allows you to authenticate with the services that are provided by the included authorities, including Quay.io, which serves the container images for OKD components.
Using a pull secret from the Red Hat OpenShift Cluster Manager is not required. You can use a pull secret for another private registry. Or, if you do not need the cluster to pull images from a private registry, you can use {"auths":{"fake":{"auth":"aWQ6cGFzcwo="}}}
as the pull secret when prompted during the installation.
If you do not use the pull secret from the Red Hat OpenShift Cluster Manager:
Red Hat Operators are not available.
The Telemetry and Insights operators do not send data to Red Hat.
Content from the Red Hat Container Catalog registry, such as image streams and Operators, are not available.
During an OKD installation, you can provide an SSH public key to the installation program. The key is passed to the Fedora CoreOS (FCOS) nodes through their Ignition config files and is used to authenticate SSH access to the nodes. The key is added to the ~/.ssh/authorized_keys
list for the core
user on each node, which enables password-less authentication.
After the key is passed to the nodes, you can use the key pair to SSH in to the FCOS nodes as the user core
. To access the nodes through SSH, the private key identity must be managed by SSH for your local user.
If you want to SSH in to your cluster nodes to perform installation debugging or disaster recovery, you must provide the SSH public key during the installation process. The ./openshift-install gather
command also requires the SSH public key to be in place on the cluster nodes.
Do not skip this procedure in production environments, where disaster recovery and debugging is required. |
You must use a local key, not one that you configured with platform-specific approaches such as AWS key pairs. |
On clusters running Fedora CoreOS (FCOS), the SSH keys specified in the Ignition config files are written to the |
If you do not have an existing SSH key pair on your local machine to use for authentication onto your cluster nodes, create one. For example, on a computer that uses a Linux operating system, run the following command:
$ ssh-keygen -t ed25519 -N '' -f <path>/<file_name> (1)
1 | Specify the path and file name, such as ~/.ssh/id_ed25519 , of the new SSH key. If you have an existing key pair, ensure your public key is in the your ~/.ssh directory. |
If you plan to install an OKD cluster that uses FIPS Validated / Modules in Process cryptographic libraries on the |
View the public SSH key:
$ cat <path>/<file_name>.pub
For example, run the following to view the ~/.ssh/id_ed25519.pub
public key:
$ cat ~/.ssh/id_ed25519.pub
Add the SSH private key identity to the SSH agent for your local user, if it has not already been added. SSH agent management of the key is required for password-less SSH authentication onto your cluster nodes, or if you want to use the ./openshift-install gather
command.
On some distributions, default SSH private key identities such as |
If the ssh-agent
process is not already running for your local user, start it as a background task:
$ eval "$(ssh-agent -s)"
Agent pid 31874
If your cluster is in FIPS mode, only use FIPS-compliant algorithms to generate the SSH key. The key must be either RSA or ECDSA. |
Add your SSH private key to the ssh-agent
:
$ ssh-add <path>/<file_name> (1)
1 | Specify the path and file name for your SSH private key, such as ~/.ssh/id_ed25519 |
Identity added: /home/<you>/<path>/<file_name> (<computer_name>)
When you install OKD, provide the SSH public key to the installation program.
The OKD installation program requires that a Fedora CoreOS (FCOS) image be present in the OpenStack cluster. Retrieve the latest FCOS image, then upload it using the OpenStack CLI.
The OpenStack CLI is installed.
Log in to the Red Hat Customer Portal’s Product Downloads page.
Under Version, select the most recent release of OKD 4.9 for Fedora 8.
The FCOS images might not change with every release of OKD. You must download images with the highest version that is less than or equal to the OKD version that you install. Use the image versions that match your OKD version if they are available. |
Download the Fedora CoreOS (FCOS) - OpenStack Image (QCOW).
Decompress the image.
You must decompress the OpenStack image before the cluster can use it. The name of the downloaded file might not contain a compression extension, like
|
From the image that you downloaded, create an image that is named rhcos
in your cluster by using the OpenStack CLI:
$ openstack image create --container-format=bare --disk-format=qcow2 --file rhcos-${RHCOS_VERSION}-openstack.qcow2 rhcos
Depending on your OpenStack environment, you might be able to upload the image in either .raw or .qcow2 formats. If you use Ceph, you must use the .raw format.
|
If the installation program finds multiple images with the same name, it chooses one of them at random. To avoid this behavior, create unique names for resources in OpenStack. |
After you upload the image to OpenStack, it is usable in the installation process.
The OKD installation process requires external network access. You must provide an external network value to it, or deployment fails. Before you begin the process, verify that a network with the external router type exists in OpenStack.
Using the OpenStack CLI, verify the name and ID of the 'External' network:
$ openstack network list --long -c ID -c Name -c "Router Type"
+--------------------------------------+----------------+-------------+
| ID | Name | Router Type |
+--------------------------------------+----------------+-------------+
| 148a8023-62a7-4672-b018-003462f8d7dc | public_network | External |
+--------------------------------------+----------------+-------------+
A network with an external router type appears in the network list. If at least one does not, see Creating a default floating IP network and Creating a default provider network.
If the Neutron trunk service plugin is enabled, a trunk port is created by default. For more information, see Neutron trunk port. |
At deployment, all OKD machines are created in a OpenStack-tenant network. Therefore, they are not accessible directly in most OpenStack deployments.
You can configure OKD API and application access by using floating IP addresses (FIPs) during installation. You can also complete an installation without configuring FIPs, but the installer will not configure a way to reach the API or applications externally.
Create floating IP (FIP) addresses for external access to the OKD API, cluster applications, and the bootstrap process.
Using the OpenStack CLI, create the API FIP:
$ openstack floating ip create --description "API <cluster_name>.<base_domain>" <external_network>
Using the OpenStack CLI, create the apps, or Ingress, FIP:
$ openstack floating ip create --description "Ingress <cluster_name>.<base_domain>" <external_network>
By using the OpenStack CLI, create the bootstrap FIP:
$ openstack floating ip create --description "bootstrap machine" <external_network>
Add records that follow these patterns to your DNS server for the API and Ingress FIPs:
api.<cluster_name>.<base_domain>. IN A <API_FIP>
*.apps.<cluster_name>.<base_domain>. IN A <apps_FIP>
If you do not control the DNS server, you can access the cluster by adding the cluster domain names such as the following to your
The cluster domain names in the |
Add the FIPs to the
inventory.yaml
file as the values of the following
variables:
os_api_fip
os_bootstrap_fip
os_ingress_fip
If you use these values, you must also enter an external network as the value of the
os_external_network
variable in the inventory.yaml
file.
You can make OKD resources available outside of the cluster by assigning a floating IP address and updating your firewall configuration. |
You can install OKD on OpenStack without providing floating IP addresses.
In the
inventory.yaml
file, do not define the following
variables:
os_api_fip
os_bootstrap_fip
os_ingress_fip
If you cannot provide an external network, you can also leave os_external_network
blank. If you do not provide a value for os_external_network
, a router is not created for you, and, without additional action, the installer will fail to retrieve an image from Glance. Later in the installation process, when you create network resources, you must configure external connectivity on your own.
If you run the installer
with the wait-for
command
from a system that cannot reach the cluster API due to a lack of floating IP addresses or name resolution, installation fails. To prevent installation failure in these cases, you can use a proxy network or run the installer from a system that is on the same network as your machines.
You can enable name resolution by creating DNS records for the API and Ingress ports. For example:
If you do not control the DNS server, you can add the record to your |
The OKD installation program relies on a file that is called clouds.yaml
. The file describes OpenStack configuration parameters, including the project name, log in information, and authorization service URLs.
Create the clouds.yaml
file:
If your OpenStack distribution includes the Horizon web UI, generate a clouds.yaml
file in it.
Remember to add a password to the |
If your OpenStack distribution does not include the Horizon web UI, or you do not want to use Horizon, create the file yourself. For detailed information about clouds.yaml
, see Config files in the OpenStack documentation.
clouds:
shiftstack:
auth:
auth_url: http://10.10.14.42:5000/v3
project_name: shiftstack
username: shiftstack_user
password: XXX
user_domain_name: Default
project_domain_name: Default
dev-env:
region_name: RegionOne
auth:
username: 'devuser'
password: XXX
project_name: 'devonly'
auth_url: 'https://10.10.14.22:5001/v2.0'
If your OpenStack installation uses self-signed certificate authority (CA) certificates for endpoint authentication:
Copy the certificate authority file to your machine.
Add the cacerts
key to the clouds.yaml
file. The value must be an absolute, non-root-accessible path to the CA certificate:
clouds:
shiftstack:
...
cacert: "/etc/pki/ca-trust/source/anchors/ca.crt.pem"
After you run the installer with a custom CA certificate, you can update the certificate by editing the value of the
|
Place the clouds.yaml
file in one of the following locations:
The value of the OS_CLIENT_CONFIG_FILE
environment variable
The current directory
A Unix-specific user configuration directory, for example ~/.config/openstack/clouds.yaml
A Unix-specific site configuration directory, for example /etc/openstack/clouds.yaml
The installation program searches for clouds.yaml
in that order.
You can customize the OKD cluster you install on OpenStack.
Obtain the OKD installation program and the pull secret for your cluster.
Obtain service principal permissions at the subscription level.
Create the install-config.yaml
file.
Change to the directory that contains the installation program and run the following command:
$ ./openshift-install create install-config --dir <installation_directory> (1)
1 | For <installation_directory> , specify the directory name to store the
files that the installation program creates. |
Specify an empty directory. Some installation assets, like bootstrap X.509 certificates have short expiration intervals, so you must not reuse an installation directory. If you want to reuse individual files from another cluster installation, you can copy them into your directory. However, the file names for the installation assets might change between releases. Use caution when copying installation files from an earlier OKD version. |
At the prompts, provide the configuration details for your cloud:
Optional: Select an SSH key to use to access your cluster machines.
For production OKD clusters on which you want to perform installation debugging or disaster recovery, specify an SSH key that your |
Select openstack as the platform to target.
Specify the OpenStack external network name to use for installing the cluster.
Specify the floating IP address to use for external access to the OpenShift API.
Specify a OpenStack flavor with at least 16 GB RAM to use for control plane nodes and 8 GB RAM for compute nodes.
Select the base domain to deploy the cluster to. All DNS records will be sub-domains of this base and will also include the cluster name.
Enter a name for your cluster. The name must be 14 or fewer characters long.
Paste the pull secret from the Red Hat OpenShift Cluster Manager. This field is optional.
Modify the install-config.yaml
file. You can find more information about
the available parameters in the "Installation configuration parameters" section.
Back up the install-config.yaml
file so that you can use
it to install multiple clusters.
The |
You now have the file install-config.yaml
in the directory that you specified.
Before you deploy an OKD cluster, you provide parameter values to describe your account on the cloud platform that hosts your cluster and optionally customize your cluster’s platform. When you create the install-config.yaml
installation configuration file, you provide values for the required parameters through the command line. If you customize your cluster, you can modify the install-config.yaml
file to provide more details about the platform.
After installation, you cannot modify these parameters in the |
The |
Required installation configuration parameters are described in the following table:
Parameter | Description | Values |
---|---|---|
|
The API version for the |
String |
|
The base domain of your cloud provider. The base domain is used to create routes to your OKD cluster components. The full DNS name for your cluster is a combination of the |
A fully-qualified domain or subdomain name, such as |
|
Kubernetes resource |
Object |
|
The name of the cluster. DNS records for the cluster are all subdomains of |
String of lowercase letters, hyphens ( |
|
The configuration for the specific platform upon which to perform the installation: |
Object |
You can customize your installation configuration based on the requirements of your existing network infrastructure. For example, you can expand the IP address block for the cluster network or provide different IP address blocks than the defaults.
Only IPv4 addresses are supported.
Parameter | Description | Values | ||
---|---|---|---|---|
|
The configuration for the cluster network. |
Object
|
||
|
The cluster network provider Container Network Interface (CNI) plugin to install. |
Either |
||
|
The IP address blocks for pods. The default value is If you specify multiple IP address blocks, the blocks must not overlap. |
An array of objects. For example:
|
||
|
Required if you use An IPv4 network. |
An IP address block in Classless Inter-Domain Routing (CIDR) notation.
The prefix length for an IPv4 block is between |
||
|
The subnet prefix length to assign to each individual node. For example, if |
A subnet prefix. The default value is |
||
|
The IP address block for services. The default value is The OpenShift SDN and OVN-Kubernetes network providers support only a single IP address block for the service network. |
An array with an IP address block in CIDR format. For example:
|
||
|
The IP address blocks for machines. If you specify multiple IP address blocks, the blocks must not overlap. |
An array of objects. For example:
|
||
|
Required if you use |
An IP network block in CIDR notation. For example,
|
Optional installation configuration parameters are described in the following table:
Parameter | Description | Values | ||||
---|---|---|---|---|---|---|
|
A PEM-encoded X.509 certificate bundle that is added to the nodes' trusted certificate store. This trust bundle may also be used when a proxy has been configured. |
String |
||||
|
The configuration for the machines that comprise the compute nodes. |
Array of |
||||
|
Determines the instruction set architecture of the machines in the pool. Currently, heteregeneous clusters are not supported, so all pools must specify the same architecture. Valid values are |
String |
||||
|
Whether to enable or disable simultaneous multithreading, or
|
|
||||
|
Required if you use |
|
||||
|
Required if you use |
|
||||
|
The number of compute machines, which are also known as worker machines, to provision. |
A positive integer greater than or equal to |
||||
|
The configuration for the machines that comprise the control plane. |
Array of |
||||
|
Determines the instruction set architecture of the machines in the pool. Currently, heterogeneous clusters are not supported, so all pools must specify the same architecture. Valid values are |
String |
||||
|
Whether to enable or disable simultaneous multithreading, or
|
|
||||
|
Required if you use |
|
||||
|
Required if you use |
|
||||
|
The number of control plane machines to provision. |
The only supported value is |
||||
|
The Cloud Credential Operator (CCO) mode. If no mode is specified, the CCO dynamically tries to determine the capabilities of the provided credentials, with a preference for mint mode on the platforms where multiple modes are supported.
|
|
||||
|
Sources and repositories for the release-image content. |
Array of objects. Includes a |
||||
|
Required if you use |
String |
||||
|
Specify one or more repositories that may also contain the same images. |
Array of strings |
||||
|
How to publish or expose the user-facing endpoints of your cluster, such as the Kubernetes API, OpenShift routes. |
Setting this field to |
||||
|
The SSH key or keys to authenticate access your cluster machines.
|
One or more keys. For example:
|
Additional OpenStack configuration parameters are described in the following table:
Parameter | Description | Values |
---|---|---|
|
For compute machines, the size in gigabytes of the root volume. If you do not set this value, machines use ephemeral storage. |
Integer, for example |
|
For compute machines, the root volume’s type. |
String, for example |
|
For control plane machines, the size in gigabytes of the root volume. If you do not set this value, machines use ephemeral storage. |
Integer, for example |
|
For control plane machines, the root volume’s type. |
String, for example |
|
The name of the OpenStack cloud to use from the list of clouds in the
|
String, for example |
|
The OpenStack external network name to be used for installation. |
String, for example |
|
The OpenStack flavor to use for control plane and compute machines. This property is deprecated. To use a flavor as the default for all machine pools, add it as the value of the |
String, for example |
Optional OpenStack configuration parameters are described in the following table:
Parameter | Description | Values |
---|---|---|
|
Additional networks that are associated with compute machines. Allowed address pairs are not created for additional networks. |
A list of one or more UUIDs as strings. For example, |
|
Additional security groups that are associated with compute machines. |
A list of one or more UUIDs as strings. For example, |
|
OpenStack Compute (Nova) availability zones (AZs) to install machines on. If this parameter is not set, the installer relies on the default settings for Nova that the OpenStack administrator configured. On clusters that use Kuryr, OpenStack Octavia does not support availability zones. Load balancers and, if you are using the Amphora provider driver, OKD services that rely on Amphora VMs, are not created according to the value of this property. |
A list of strings. For example, |
|
For compute machines, the availability zone to install root volumes on. If you do not set a value for this parameter, the installer selects the default availability zone. |
A list of strings, for example |
|
Additional networks that are associated with control plane machines. Allowed address pairs are not created for additional networks. |
A list of one or more UUIDs as strings. For example, |
|
Additional security groups that are associated with control plane machines. |
A list of one or more UUIDs as strings. For example, |
|
OpenStack Compute (Nova) availability zones (AZs) to install machines on. If this parameter is not set, the installer relies on the default settings for Nova that the OpenStack administrator configured. On clusters that use Kuryr, OpenStack Octavia does not support availability zones. Load balancers and, if you are using the Amphora provider driver, OKD services that rely on Amphora VMs, are not created according to the value of this property. |
A list of strings. For example, |
|
For control plane machines, the availability zone to install root volumes on. If you do not set this value, the installer selects the default availability zone. |
A list of strings, for example |
|
The location from which the installer downloads the FCOS image. You must set this parameter to perform an installation in a restricted network. |
An HTTP or HTTPS URL, optionally with an SHA-256 checksum. For example, |
|
Properties to add to the installer-uploaded ClusterOSImage in Glance. This property is ignored if You can use this property to exceed the default persistent volume (PV) limit for OpenStack of 26 PVs per node. To exceed the limit, set the You can also use this property to enable the QEMU guest agent by including the |
A list of key-value string pairs. For example, |
|
The default machine pool platform configuration. |
|
|
An existing floating IP address to associate with the Ingress port. To use this property, you must also define the |
An IP address, for example |
|
An existing floating IP address to associate with the API load balancer. To use this property, you must also define the |
An IP address, for example |
|
IP addresses for external DNS servers that cluster instances use for DNS resolution. |
A list of IP addresses as strings. For example, |
|
The UUID of a OpenStack subnet that the cluster’s nodes use. Nodes and virtual IP (VIP) ports are created on this subnet. The first item in If you deploy to a custom subnet, you cannot specify an external DNS server to the OKD installer. Instead, add DNS to the subnet in OpenStack. |
A UUID as a string. For example, |
Optionally, you can deploy a cluster on a OpenStack subnet of your choice. The subnet’s GUID is passed as the value of platform.openstack.machinesSubnet
in the install-config.yaml
file.
This subnet is used as the cluster’s primary subnet. By default, nodes and ports are created on it. You can create nodes and ports on a different OpenStack subnet by setting the value of the platform.openstack.machinesSubnet
property to the subnet’s UUID.
Before you run the OKD installer with a custom subnet, verify that your configuration meets the following requirements:
The subnet that is used by platform.openstack.machinesSubnet
has DHCP enabled.
The CIDR of platform.openstack.machinesSubnet
matches the CIDR of networking.machineNetwork
.
The installation program user has permission to create ports on this network, including ports with fixed IP addresses.
Clusters that use custom subnets have the following limitations:
If you plan to install a cluster that uses floating IP addresses, the platform.openstack.machinesSubnet
subnet must be attached to a router that is connected to the externalNetwork
network.
If the platform.openstack.machinesSubnet
value is set in the install-config.yaml
file, the installation program does not create a private network or subnet for your OpenStack machines.
You cannot use the platform.openstack.externalDNS
property at the same time as a custom subnet. To add DNS to a cluster that uses a custom subnet, configure DNS on the OpenStack network.
By default, the API VIP takes x.x.x.5 and the Ingress VIP takes x.x.x.7 from your network’s CIDR block. To override these default values,
set values for |
install-config.yaml
file for OpenStack with KuryrTo deploy with Kuryr SDN instead of the default OpenShift SDN, you must
modify the install-config.yaml
file to include Kuryr
as the desired
networking.networkType
and proceed with the default OKD SDN installation steps.
This sample install-config.yaml
demonstrates all of the possible
OpenStack customization options.
This sample file is provided for reference only. You must obtain your
|
apiVersion: v1
baseDomain: example.com
controlPlane:
name: master
platform: {}
replicas: 3
compute:
- name: worker
platform:
openstack:
type: ml.large
replicas: 3
metadata:
name: example
networking:
clusterNetwork:
- cidr: 10.128.0.0/14
hostPrefix: 23
machineNetwork:
- cidr: 10.0.0.0/16
serviceNetwork:
- 172.30.0.0/16 (1)
networkType: Kuryr
platform:
openstack:
cloud: mycloud
externalNetwork: external
computeFlavor: m1.xlarge
apiFloatingIP: 128.0.0.1
trunkSupport: true (2)
octaviaSupport: true (2)
pullSecret: '{"auths": ...}'
sshKey: ssh-ed25519 AAAA...
1 | The Amphora Octavia driver creates two ports per load balancer. As a
result, the service subnet that the installer creates is twice the size of the
CIDR that is specified as the value of the serviceNetwork property. The larger range is
required to prevent IP address conflicts. |
2 | Both trunkSupport and octaviaSupport are automatically discovered by the
installer, so there is no need to set them. But if your environment does not
meet both requirements, Kuryr SDN will not properly work. Trunks are needed
to connect the pods to the OpenStack network and Octavia is required to create the
OKD services. |
You can deploy your OKD clusters on OpenStack with a primary network interface on a provider network. Provider networks are commonly used to give projects direct access to a public network that can be used to reach the internet. You can also share provider networks among projects as part of the network creation process.
OpenStack provider networks map directly to an existing physical network in the data center. A OpenStack administrator must create them.
In the following example, OKD workloads are connected to a data center by using a provider network: