Crunchy Postgres for Kubernetes 5.8.x Release notes

Release notes for each of the 5.8.x releases.

Component versions

Postgres extension versions

A bold version number indicates that the component version was updated in latest release.

5.8.2

Breaking Changes

• OTel Collector is now at 0.125.0. This update is incompatible with previous CPK versions. If you are using the OpenTelemetryLogging or OpenTelemetryMetrics feature gates, you will need to update the OTel Collector sidecar. See the Configuring Cluster Images documentation for additional guidance updating your image.

Changes

• PostgreSQL versions 17.5, 16.9, 15.13, 14.18 and 13.21 are now available.
• The pgBouncer is now at version 1.24.1.
• The Citus 13.0.3 is now available for Postgres 15, 16 and 17.
• The orafce is now at version 4.14.2.
• The pg_parquet 0.4.0 extenion is now available.
• The Timescaledb extension is at version 2.19.3 for PG 17, 16, 15, and 14.

Fixes

• PostgresCluster now allows the --target-timeline restore option.
• OpenTelemetry Logs and Metrics can now be turned on with backups turned off.
• Metrics for replication slots now included with OpenTelemetryMetrics feature.
• Removed fieldPath CEL validation to maintain support for OpenShift v4.14.
• Metric and functionality for pg_hba_checksum now included with OpenTelemetryMetrics feature.
• OpenTelemetryMetrics: the ccp_stat_database metric is now collected every 5 seconds, and the ccp_stat_user_tables metric is now collected every 5 minutes.
• pgBackRest repository host pods no longer enter a CrashLoopBackOff when creating pgBackRest directories on CIFS mounts.

5.8.1

Changes

• Postgres Exporter is now at 0.17.1.
• pgAdmin is now at version 9.2.
• The Timescaledb extension is at version 2.18.2 for PG 17, 16, 15, and 14.

5.8.0

Features

• A new API is available for deploying and managing Crunchy Data Warehouse. Please note that deploying Crunchy Data Warehouse requires access to the Crunchy Data Warehouse container image, which is available by commercial subscription only at this time. For more information, contact Crunchy Data at info@crunchydata.com.
• Enable OpenTelemetry logging for your Postgres cluster, allowing you to export logging information for your databases and supporting infrastructure (specifically high availability, disaster recovery, connection pooling and pgAdmin logs) to a variety of OpenTelemetry compatible services and backends. Use the OpenTelemetryLogging feature gate to enable OpenTelemetry logging.
• Enable OpenTelemetry metrics collection as a seamless alternative to using the Postgres Exporter for metrics collection. Use the OpenTelemetryMetrics feature gate to enable OpenTelemetry metrics.
• Collect and visualize PgBouncer metrics using OpenTelemetry metrics collection and the CPK monitoring stack. The OpenTelemetryMetrics feature gate must be enabled to collect PgBouncer metrics.
• Operator metrics created by controller-runtime are now scraped by Prometheus when the Crunchy Postgres for Kubernetes monitoring stack is deployed.
• You can store Patroni logs in the Postgres data volume by configuring spec.patroni.logging in the PostgresCluster spec.
• Starting with Crunchy Postgres for Kubernetes v5.8, all containers are built using the Red Hat Universal Base Image 9 (UBI 9).
• Configure Postgres client authentication settings and configuration parameters using the spec.authentication and spec.config.parameters fields in PostgresCluster. These provide immediate feedback about ignored and acceptable values.
• Use Secrets to configure OAuth2 authentication in PGAdmin.
• Provide ipFamilies and ipFamilyPolicy settings when configuring Services for the Postgres database and PgBouncer.
• Customize the credentials for the PgBouncer system account by updating the password field in the PgBouncer Secret.
• The operator emits a warning event when a PostgresCluster is using a major version of Postgres that is no longer receiving updates.

Breaking Changes

• Component images are now tagged with a consistent suffix. Images with the same suffix are compatible with each other. See the Container Tags section of the documentation for more details.

• PostGIS 3.0 is not supported in Crunchy Postgres for Kubernetes v5.8, and UBI 9 based builds of PostGIS 3.0 will not be provided.

• UBI 9 builds of pgAdmin4 v4.30 will not be included with Crunchy Postgres for Kubernetes v5.8. While UBI 8 builds of pgAdmin v4.30 will remain compatible with Crunchy Postgres for Kubernetes v5.8, please migrate to the PGAdmin API to ensure you are using the latest version of pgAdmin available.

• UBI 9 provides faster sorting for builtin collations, but also requires you to recreate indexes in your existing databases. Follow the instructions in Changing Base Images when upgrading to UBI 9.

  Please note that you do not need to immediately update your Postgres and component container images to UBI 9 following the upgrade to CPK v5.8. This means UBI 8 containers will remain compatible for the time being, so that you can update your PostgresCluster's at your earliest convenience. You can use related images to run Crunchy Postgres for Kubernetes v5.8 while still using UBI 8 images for the various PostgresCluster containers. It is recommended that you update your images as soon as possible to ensure you are using Postgres and component containers that are fully validated with the CPK v5.8 release line.

Changes

• The InstanceSidecars feature gate is now enabled by default.
• Validation has been added to ensure only cloud-based repos can be used with spec.datasource.pgbackrest.
• MD5 authentication is now disabled by default. Managed users have always used SCRAM-SHA-256 and are unaffected. If you have set custom MD5 passwords, use authentication rules to re-enable their access.
• PgAdmin4 is now at version 9.1.
• PgBouncer is now at version 1.24.0.
• PgBackRest is now at version 2.54.2.

Fixes

• Addressed an edge case where a snapshot doesn’t yet have a status field, which resulted in a nil pointer error.
• Changed the pgAdmin readinessProbe path to ensure proper cleanup of pgAdmin session files. Note: After upgrading an existing pgAdmin pod, it is safe to manually delete any old session files that are not automatically deleted by the cleanup script.
• Standby clusters now shut down properly when spec.shutdown is set to true.