Upgrades and removal

Upgrade

Upgrade to a release with the same values. Use --reset-then-reuse-values: --reuse-values keeps the old chart's defaults.

helm upgrade relic oci://ghcr.io/relicbuild/charts/relic --version 0.2.0-beta.4 -n relic --reset-then-reuse-values

• Migrations are forward-only. The server applies them before it reports ready, within the five-minute startup probe.
• Downgrades are not supported; restore from a backup instead.
• On shutdown, the server finishes in-flight requests for up to 110 seconds (grace period 120 seconds, 450 on a single node).

Each release's chart pins its dependencies; the release notes say when one changes.

Helm never upgrades CRDs. Before you upgrade to a release with a new CloudNativePG operator, download its chart:

helm pull oci://ghcr.io/relicbuild/charts/relic --version 0.2.0-beta.4 --untar

Apply the chart's copy of the CRDs:

kubectl apply --server-side -f relic/crds/cloudnative-pg.yaml

For the Prometheus Operator's CRDs, see Monitoring.

Uninstall

To keep the data, take a backup and keep the encryption key.

Remove the release:

helm uninstall relic -n relic

These remain, and a new install in the namespace reuses them:

• Volumes that StatefulSets created, the single-node data volume and the vulnerability database volume.
• The credentials Secret, relic-credentials, with the encryption key.
• The SeaweedFS S3 Secret.

To start from nothing, delete them. List the volume claims:

kubectl -n relic get pvc

Delete all volume claims in the namespace:

kubectl -n relic delete pvc --all

Delete the credentials Secret:

kubectl -n relic delete secret relic-credentials

Delete the SeaweedFS S3 Secret:

kubectl -n relic delete secret relic-seaweedfs-s3-secret