Tanzu SQL with Postgres for Kubernetes
This document contains pertinent release information about VMware Tanzu SQL with Postges for Kubernetes. Obtain the most recent version of the distribution from VMware Tanzu Network.
Supported Platforms
Tanzu SQL with Postgres versions 1.X are supported on the following platforms:
- VMware Tanzu Kubernetes Grid Integrated Edition (TKGi)
- Google Kubernetes Engine (GKE) Kubernetes
Additional Kubernetes environments, such as Minikube, can be used for testing or demonstration purposes.
Release 1.1.0
Release Date: February 26, 2021
Software Components
VMware Postgres Version | Component | Component Version |
---|---|---|
1.1.0 | PostgreSQL | 11.10 |
pgBackRest | 2.31 | |
pg_auto_failover | 1.4.0 |
Features
Tanzu SQL with Postgres 1.1.0 has the following features:
- Support for upgrading from Tanzu Postgres 1.0.0 to 1.1.0. See Upgrading the Tanzu Postgres Operator and Instances.
- Support for Postgres 11.10.
- Enhanced security by implementing Postgres cluster communications via SSL.
- Improved auto-healing, when instances or services are terminated abnormally or accidentally. The Postgres operator monitors and automatically restarts any deleted or stopped instances or agents.
- Postgres instances with the same name, in different namespaces, can now be backed up to the same S3 location.
Changed Features
- Updated the pgbackrest sample configuration file, from
pgbackrest.conf
topgbackrest.conf.template
. - The
storageSize
parameter cannot be altered after Postgres instance creation. Any attempt to do so generates an error similar to:storageSize cannot be reduced after the instance is created. No changes have been made to the running instance.
Fixed Issues
- [166560384] - Tanzu Postgres backups to an S3 location, using the parameter
verifyTLS: true
and a well-known Certificate Authority, would fail with an error similar to:2020-12-18 01:01:43.460 P00 DEBUG: common/io/http/request::httpRequestProcess: retry CryptoError: unable to verify certificate presented by 's3.us-west-1.amazonaws.com:443': [20] unable to get local issuer certificate
This issue has been resolved. - [175791284] - Fixed an issue where
storageSize
updates affected more than one instance, if the instances had matching names in separate namespaces. - [175885808] - Updated the
s3-secret-example.yaml
file for S3 backups, and all parameters are now specified in double quotes. - [175618701] - Resolved an issue with the s3 secret
yaml
file, where parameters marked as “optional” but not configured would cause the backup operation to fail. - [175602831] - When the
archive_mode
flag wason
(in thepostgresql.conf
file), but thebackupLocationSecret
was left unconfigured in the instance configurationyaml
file, backups were still attempted. This issue has been resolved. - [176061339] - Changed the detail level of the pgbackrest console output to
info
.
Release 1.0.0
Release Date: October 30, 2020
Tanzu SQL with Postgres 1.0.0 is the first release of VMware Tanzu SQL with Postgres on Kubernetes.
Software Components
VMware Postgres Version | Component | Component Version |
---|---|---|
1.0.0 | PostgreSQL | 11.9 |
psqlODBC | 11.0-0000 | |
pgjdbc | 42.2.5 | |
pgBackRest | 2.28 | |
pg_auto_failover | 1.4.0 | |
postGIS | 2.5.4 |
Features
Tanzu SQL with Postgres 1.0.0 has the following features:
- Support for backing up the Postgres instances to an S3 compatible storage location. See Backing Up and Restoring Tanzu Postgres.
- Support for creating a High Availability cluster configuration, with one primary and a mirror. See Configuring High Availability in Tanzu Postgres .
- Flexibility to update the Postgres images after deployment, and reconfigure CPU, memory, and persistent volume storage sizes. See Updating a Postgres Instance Configuration for more information.
Known Issues and Limitations
- Upgrades from the Beta program releases to the Tanzu SQL with Postgres 1.0.0 release are not supported. Download and install the latest version.
- The High Availability configuration contains only one mirror.
- The default storage size for the Postgres instance is too limited for long term running Postgres environments. Change the
storageSize
to 10G, and use an expandable storage class. See Configuring a Postgres Instance. - During an upgrade from 1.0.0 to 1.1.0, in an HA scenario, the Postgres instances state does not show “Ready” until both the primary and the mirror nodes have restarted. This limitation stops the clients from connecting to a read-write instance during the upgrade.