Skip to main content
Version: Next

Upgrade OpenEverest

The OpenEverest core and each provider are separate Helm releases with their own versions. You upgrade them separately: first the core, then the providers.

Developer Preview

Upgrades between OpenEverest 2 Developer Preview releases are not supported. The Instance and Provider APIs change between previews without a migration path, so install each preview on a fresh cluster. Upgrading from OpenEverest 1.x to 2 is not supported either; a migration path is planned for the General Availability release.

The procedure below describes how upgrades work, for patch releases and for future releases.

Before you upgrade​

  • Read the release notes of the target version, especially the breaking changes.

  • Check which providers support the target version. The Plugin Hub and the provider release notes show the compatible core versions.

  • Back up your instances. See Create backups.

  • Check your current releases:

    helm list -n everest-system

1. Upgrade the custom resource definitions​

Helm does not upgrade CRDs during helm upgrade, and the OpenEverest chart refuses to upgrade until the CRDs match the target version. Upgrade them first with the everest-crds chart:

helm repo update
helm upgrade --install everest-crds openeverest/everest-crds \
  --namespace everest-system \
  --version <version> \
  --take-ownership

--take-ownership requires Helm 3.17 or later. It lets the everest-crds release adopt CRDs that were installed by the openeverest chart.

2. Upgrade the core​

helm upgrade everest-core openeverest/openeverest \
  --namespace everest-system \
  --version <version> \
  --reuse-values
tip

Without --reuse-values, Helm resets every value you don't pass on the command line to the chart default. Either use --reuse-values or pass your complete values file with -f.

The chart runs pre-upgrade checks. To skip them, set upgrade.preflightChecks=false and upgrade.crdChecks=false — only do this if you know why a check fails.

3. Upgrade the providers​

Upgrade each provider to a version that supports the new core:

helm upgrade provider-percona-postgresql \
  oci://ghcr.io/openeverest/charts/provider-percona-postgresql \
  --namespace everest-system \
  --version <provider-version> \
  --reuse-values

A provider upgrade can change the components of running instances, for example new default images. Changes that would restart or disrupt an instance are applied according to the instance's maintenance settings.

4. Check the result​

kubectl get pods -n everest-system
kubectl get providers
kubectl get instances -A

Log in to the UI and check that your instances are Ready.