Pivotal Greenplum for Kubernetes 0.3.0 Release Notes

Pivotal Greenplum for Kubernetes is an experimental product and is not intended for use in a production environment. Experimental features are subject to change without notice in future releases.

Pivotal Greenplum for Kubernetes is an experimental release of Pivotal Greenplum Database on Kubernetes. Greenplum for Kubernetes 0.3.0 is a minor release that includes several new features, feature changes, and bug fixes; it is based on Pivotal Greenplum Database 5.11.3.

Refer to the Pivotal Greenplum Database documentation for detailed information on Pivotal Greenplum Database.

Scope

Pivotal Greenplum for Kubernetes 0.3.0 deploys on PKS version 1.1.1 or later or Kubernetes version 1.10.4 or later (Minikube, GKE, and so forth).

Components

Pivotal Greenplum for Kubernetes includes the Greenplum Database version identified below:

Greenplum for Kubernetes Version Greenplum Database Version
0.3.0 5.11.3
0.2.0 5.11.0
0.1.0 5.10.0

Changes from Previous Release

This release introduces the following feature changes and improvements:

  • The Greenplum Operator has been improved to provide for a wide range of deployment options, including CPU and memory limits. All customers who are evaluating this experimental release of Greenplum for Kubernetes are encouraged to use the Greenplum Operator for test deployments. The bash scripts used for previous Kubernetes deployments are deprecated, and will be removed in a future release.
  • Each Greenplum instance uses a configured Storage Class resource in the Kubernetes cluster to dynamically provision PersistenVolumes for storage. You can configure the Storage Class according to your performance needs. See Storage Classes in the Kubernetes documentation to understand the different configuration options. The default Greenplum Operator manifest targets a Storage Class named “standard”, which is the default name used for most Kubernetes configurations. Pivotal recommends using a Storage Class that is provisioned with a local SSD and XFS filesystem for best performance.
  • This release includes support for QuickLZ compression in Greenplum tables. See Choosing the Table Storage Model in the Pivotal Greenplum Database documentation for more information.
  • Most Greenplum instances that you deploy using the Greenplum Operator are labeled with the Greenplum cluster name, so you can easily manage them with kubectl commands (for example, kubectl get all -l greenplum-cluster=my-greenplum.)

Known Issues and Limitations

  • Pivotal Greenplum for Kubernetes is an experimental product and is not intended for use in a production environment. Experimental features are subject to change without notice in future releases.
  • The Greenplum Operator does not yet support changing attribute values of a deployed Greenplum cluster. In order to change attributes such as the number of segments in the cluster, the cluster name, or segment resource usage, you must delete the cluster and then redeploy using the changed values.
  • In order to install Greenplum Database package (.gppkg) files, such as MADLib, with the gppkg utility, you must first install the alien utility onto Greenplum hosts (for example, sudo apt-get install alien).
  • Certain Greenplum resouces that are created during deployment are not yet labeled for convenient management with kubectl. This includes Greenplum secret objects, ConfigMap, Services.