Skip to content

Latest commit

 

History

History
60 lines (35 loc) · 4.02 KB

upgrades.md

File metadata and controls

60 lines (35 loc) · 4.02 KB
Raw

Upgrade - 0.5.1 to 1.0.0

Release 1.0.0 includes a backward incompatible change from version 0.5.1 as we have moved from a Go based operator to a Helm based operator.

OLM upgrade - 0.5.1 to 1.0.0

Note: The nginx-ingress-operator supports Basic Install only - we do not support auto-updates. When you are installing the Operator using the OLM, the auto-update feature should be disabled to avoid breaking changes being auto-applied. In OpenShift, this can be done by setting the Approval Strategy to Manual. Please see the Operator SDK docs for more details on the Operator Capability Levels.

  1. Upgrade CRDs
  2. Uninstall Go operator -> this will also remove any instances of the NginxIngressController, but not any dependent objects (ingresses, VSs, etc)
  3. Remove the nginx-ingress ingressClass kubectl delete ingressclass/nginx
  4. Install new operator
  5. Deploy common resources (scc, default server secret, ns, etc). Note: Multiple NIC deployments: the RBAC resources should be deployed separately if deploying multiple ICs in same namespace. This is because only one of the ICs in a namespace will be assigned "ownership" of these resources. Similarly, the IngressClass resource needs to be created separately if deploying multiple NIC instances with a shared IngressClass. See the README for more information
  6. Re-create ingress controllers (note: multi IC rules) using the new Operator. Be sure to use the same configuration as the previous deployments (ingress class name, namespaces etc). They will pick up all deployed dependent resources.

1. Upgrade the existing NIC crds

Navigate here and run kubectl apply -f crds/

2. Uninstall the existing 0.5.1 operator, the nginx ingress controller CRD, and the ingressClass

Uninstall the operator using the web console - see the OCP documentation for details.

Next uninstall the NIC CRD nginxingresscontrollers.k8s.nginx.org. This will remove any instances of the NginxIngressController, but not any dependent objects (ingresses, VSs, etc).

Finally, remove the nginx-ingress ingressClass kubectl delete ingressclass/nginx.

3. Install the latest version of the operator

Install the latest version of the Operator following the steps outlined in OpenShift installation doc.

4. Deploy new ingress controller deployments

Use the new Nginx Ingress Operator installation to deploy Nginx Ingress Controller - see the release notes here and a guide to the Helm configuration parameters here

Manual upgrade - 0.5.1 to 1.0.0

1. Deploy the new operator

Deploy the operator following the steps outlined in manual installation doc.

2. Cleanup the existing operator

Uninstall the existing operator deployment:

  1. Checkout the previous version of the nginx-ingress-operator 0.5.1.

  2. Uninstall the resources by running the following command:

    make undeploy

3. Install the latest version of the operator

Install the latest version of the Operator following the steps outlined in manual installation doc.

4. Deploy new ingress controller deployments

Use the new Nginx Ingress Operator installation to deploy Nginx Ingress Controller - see the release notes here and a guide to the Helm configuration parameters here