Data-bench

databench.clihelper.CLIGenerator lost the ability to read a configuration file in addition to the command-line. The idea is to reduce the number of places where something can be defined and avoid confusing the end user.

This release is a near complete re-write of all kubernetes configuration files:
- New Apache Zookeeper deployment.
- New Apache Kafka deployment.
- New Apache Spark master and worker deployments.
- Updated generator and consumer deployments.
- Container images from [data-bench-container][1] tagged with 0.4.0.
- README.md documenting each set of deployment files.
- Shell environments for each deployment via ConfigMap.
- Configuration files for each deployment via ConfigMap.
- Flattened kubernetes namespace, all pods running in the 'default' namespace.
- Removed persistent storage requirement (now optional).

Our goal for this release was to simplify deployment of pods while increasing flexibility via the use of [ConfigMaps][2]. The pod namespace was also regularized via the use of Kubernetes StatefulSets which results in more friendly pod names even though the pods aren't strictly stateful or sets, e.g.:
- cassandra-0
- kafka-0
- zookeeper-0
- spark-master-0
- spark-worker-0
- generators-0
- consumers=0

Cluster services are:
- cassandra.default.svc.cluster.local
- spark-master.default.svc.cluster.local
- kafka.default.svc.cluster.local

INCOMPATIBILTIES

All of these changes are supported by the [data-bench-containers][1] version 0.4.0 release and are incompatible with earlier container versions.

[1]: https://github.com/Data-Bench/data-bench-containers
[2]: https://kubernetes.io/docs/tasks/configure-pod-container/configure-pod-configmap/

Fixed a bug in databench.Transaction that raised an exception in the timestamp property.

Transaction timestamp units changed from milliseconds to nanoseconds.

codescan-01
First internal code scan point release.

First release of Data Bench Faker.