temporarily support spark 2.x

[songgane-zz]

Problem :
the spark metric does not display correctly.

Cause :
This is caused by the inability to parse the SHS event log based on spark 2.x.
SparkDataCollection class processes SHS event log based on spark 1.4, but spark 2.x will cause error due to newly added SHS event log.

I did :
As a temporary measure, I changed the sbt spark version to 2.2.1 or less, and when calling the ReplayListenerBus replay, I made use of the ReplayEventsFilter to support spark 2.x with the exception of the newly added SHS event log in 2.x.

[songgane-zz]

1. SparkUtilsTest error cause :
   at hadoop 2.7.3, "Cause: java.net.UnknownHostException: sample.grid.example.com" error was occurred.
   so i change to hadoop 2.3.0 (dr-elephant's default hadoop version)
2. I accidentally omitted a statement.

[songgane-zz]

Your build configuration was configured to support java 1.7 and 1.8.
To support java 1.7, I am changed the spark version to 2.1 or lower.

[songgane-zz]

It works fine in my machines.
But I'll check build configuration...

[songgane-zz]

Without much modification, I modified the travis config and tested it.
It worked fine.

language: scala
sudo: true
jdk:
  - oraclejdk8
  - oraclejdk7
python: "2.6"
install:
  - sudo pip install inspyred
  - sudo pip install argparse

# only build PRs and master (not all branch pushes)
branches:
  only:
    - feature/support_spark_2.x

[bteeuwen]

Any news on integrating dr elephant with spark 2.x? At booking.com we're using 2.3.1, but trying out dr elephant for spark job monitoring has stalled because of incompatibility issues.

[dmateusp]

Any news on this ? What's the plan to integrate with Spark 2.x in the future? Should the support be made on a different branch ?