Hi! we now have changes in ([3] and [4]) that enable developers to easily run I/O integration tests from their local machines without having a pre-existing instance of the data store that they're running the tests against - all they need is a kubernetes cluster. It's currently hooked up for jdbc and it should be straightforward to hook it up for the other IOs that have kubernetes scripts (elasticsearchIO is the one I would suggest tackling next).
The documentation is still in a PR [1], but it walks through all the details - look for the section on running I/O Integration Tests. There's also a PR out that will enable running the JdbcIOIT in jenkins [2], so we'll have CI coverage for IO ITs. In other new, I'll be out for a while (finishing up my paternity leave) and won't be checking email, but cham and jason have context on this and can answer questions. Later! Stephen [1] IO IT docs: https://github.com/apache/beam-site/pull/275 [2] jdbc jenkins re-enabled: https://github.com/apache/beam/pull/3604 [3] maven profile changes to add io-it-suite: https://github.com/apache/beam/pull/3588 [4] pkb support for running IO ITs: https://github.com/GoogleCloudPlatform/PerfKitBenchmarker/pull/1438
