[BEAM-1740] Upgrade to bigtable-client-core 0.9.5.1

[andrewsmartin]

Be sure to do all of the following to help us incorporate your contribution
quickly and easily:

• Make sure the PR title is formatted like:
  [BEAM-<Jira issue #>] Description of pull request
• Make sure tests pass via mvn clean verify. (Even better, enable
  Travis-CI on your fork and ensure the whole test matrix passes).
• Replace <Jira issue #> in the title with the actual Jira issue
  number, if there is one.
• If this contribution is large, please file an Apache
  Individual Contributor License Agreement.

---

[coveralls]

Coverage increased (+0.01%) to 70.129% when pulling 3860e39 on andrewsmartin:master into b672cde on apache:master.

[asfbot]

Refer to this link for build results (access rights to CI server needed):
https://builds.apache.org/job/beam_PreCommit_Java_MavenInstall/8521/

Failed Tests: 1

beam_PreCommit_Java_MavenInstall/org.apache.beam:beam-runners-spark: 1

• org.apache.beam.runners.spark.translation.streaming.CreateStreamTest.testMultipleStreams

--none--

[davorbonaci]

R: @dhalperi (welcome back!)

[dhalperi]

LGTM assuming the tests pass!

@tgroh -- any additional changes regarding shading?

[dhalperi]

@davorbonaci @ssisk @jasonkuster I did not see BigtableITs being run in the precommit where they used to be run. Where are they being run now?

[tgroh]

No, this should be fine. We may still not be able to shade the Google Cloud Platform IOs without additional changes as bigtable-client-core still exposes Guava on its API surface in return types (specifically in BigtableDataClient with listenable futures and ImmutableList sampleRowKeys(...)

[tgroh]

BigtableITs are run in postcommit.

here's one

[dhalperi]

So it looks like this commit breaks those tests. @andrewsmartin can you please take a look? It might be there there has been a change in the version of some Bigtable dependency (gRPC, netty, etc.) https://builds.apache.org/job/beam_PostCommit_Java_MavenInstall/org.apache.beam$beam-sdks-java-io-google-cloud-platform/2959/testReport/junit/org.apache.beam.sdk.io.gcp.bigtable/BigtableReadIT/testE2EBigtableRead/

[dhalperi]

Btw, apparently @andrewsmartin does not have privileges to run the tests without one of the commiters running it for him.

Andrew, I'm assuming you have access to your own Bigtable cluster (if you use BigtableIO) -- suggest running the IT manually in your local developer environment.

[andrewsmartin]

@dhalperi You're right, there was a version conflict with netty. This version of the bigtable client relies on a class that only exists in netty-common:4.1.6.Final. I added another commit to bump the netty version project-wide.

[dhalperi]

I kicked off https://builds.apache.org/view/Beam/job/beam_PostCommit_Java_MavenInstall/2972/ to run the ITs manually. Thanks @andrewsmartin

[coveralls]

Coverage remained the same at 70.159% when pulling 54393e1 on andrewsmartin:master into 030528f on apache:master.

[asfbot]

Refer to this link for build results (access rights to CI server needed):
https://builds.apache.org/job/beam_PreCommit_Java_MavenInstall/8632/
--none--

[dhalperi]

Passed! Merged, thanks!