[BEAM-1856]HDFSFileSink class do not use the same configuration in master thread and slave thread.

[wypb]

As described in BEAM-1856 HDFSFileSink class do not use the same configuration in master thread and slave thread.

[coveralls]

Coverage increased (+0.01%) to 70.145% when pulling 48433aa on 397090770:master into ea33e33 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/9046/

Failed Tests: 1

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

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

--none--

[coveralls]

Coverage decreased (-0.003%) to 70.132% when pulling 48433aa on 397090770:master into ea33e33 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/9047/
--none--

[davorbonaci]

This looks good to me, but perhaps @jbonofre or @ssisk have more opinions.

R: @jbonofre @ssisk

[jbonofre]

A first glance on it: it looks good. I'm testing on couple of pipeline dealing with HDFS to be double check.

[jbonofre]

LGTM, I gonna merge.