Feast Core: Stage files manually when launching Dataflow jobs

[davidheryanto]

To resolve error when launching Dataflow jobs from Feast Core:

Caused by: java.lang.IllegalArgumentException: Unable to convert url (jar:file:/opt/feast/core/target/feast-core-0.3.0-SNAPSHOT.jar!/BOOT-INF/classes!/) to file.
	at org.apache.beam.runners.core.construction.PipelineResources.detectClassPathResourcesToStage(PipelineResources.java:67)
	at org.apache.beam.runners.dataflow.DataflowRunner.fromOptions(DataflowRunner.java:285)
	... 38 common frames omitted

When Feast Core starts from a packaged jar.

Reference:
https://stackoverflow.com/questions/48961440/illegalargumentexception-unable-to-convert-url-jarfile-app-jar-boot-inf-cla

https://stackoverflow.com/questions/48292491/java-dataflow-unable-to-use-classloader-to-detect-classpath-elements

[feast-ci-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: davidheryanto

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• OWNERS [davidheryanto]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[davidheryanto]

/hold
Stil have issues when the Dataflow job is starting up.

 A JNI error has occurred, please check your installation and try again
Caused by: java.lang.ClassNotFoundException: org.apache.beam.vendor.guava.v26_0_jre.com.google.common.graph.Network

Reference:
https://stackoverflow.com/questions/40099952/launching-dataflow-jobs-from-a-java-application?rq=1

[davidheryanto]

New approach that unpack the packaged jar and stage the files manually when trying to submit Dataflow jobs from Feast Core seems to work.

But the extractJar method is too slow . Takes more than 5 minutes to unpack Feast Core jar using extractJar method in this commit.

As a comparison, using unar command takes less than 10 seconds to unpack the jar.

[davidheryanto]

Current workaround for the slow unpacking of jar when running Feast in production:
9ba7c90

[davidheryanto]

/hold cancel

[zhilingc]

/lgtm

[zhilingc]

/retest