dpebot dependency updates + test fixes.

[dzlier-gcp]

Also add "sponge_log" to end of failsafe error reports so they show up in fusion.

[lesv]

Should this be in java-repo-tools or somewhere else?

[kurtisvg]

The parent pom is a version behind, it probably just needs to be manually updated since we released a new version

[dzlier-gcp]

This is the only failsafe plugin usage in the whole project, so I wasn't sure if it was necessary to move it up. But the parent pom is the version where I added sponge_log to surefire - it's just that failsafe and surefire don't use the same plugin config.

[lesv]

I ask questions when I review the answer can be, (and in this case appears to be), NO, this is the right spot.

[dzlier-gcp]

Updating the apache beam version causes this error in the spanner tests:

Tests run: 4, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 19.525 sec <<< FAILURE! - in com.example.dataflow.SpannerReadIT
readDbEndToEnd(com.example.dataflow.SpannerReadIT)  Time elapsed: 5.692 sec  <<< ERROR!
org.apache.beam.sdk.Pipeline$PipelineExecutionException: com.google.cloud.spanner.SpannerException: INVALID_ARGUMENT: io.grpc.StatusRuntimeException: INVALID_ARGUMENT: Query is not root partitionable since it does not have a DistributedUnion at the root. Please run EXPLAIN for query plan details.
	at com.example.dataflow.SpannerReadIT.readDbEndToEnd(SpannerReadIT.java:140)
Caused by: com.google.cloud.spanner.SpannerException: INVALID_ARGUMENT: io.grpc.StatusRuntimeException: INVALID_ARGUMENT: Query is not root partitionable since it does not have a DistributedUnion at the root. Please run EXPLAIN for query plan details.
Caused by: java.util.concurrent.ExecutionException: io.grpc.StatusRuntimeException: INVALID_ARGUMENT: Query is not root partitionable since it does not have a DistributedUnion at the root. Please run EXPLAIN for query plan details.
Caused by: io.grpc.StatusRuntimeException: INVALID_ARGUMENT: Query is not root partitionable since it does not have a DistributedUnion at the root. Please run EXPLAIN for query plan details.

I can't find much information about how to ensure the code in [SpannerReadAll.java](dataflow/spanner-io/src/main/java/com/example/dataflow/SpannerReadAll.java) has a Distributed union at the root. When I try to mimic the Sql version of that in the console, it passes just fine. Is this a bug I should file with the spanner team, or should I revert the apache beam version?

[lesv]

You might wish to make sure that beam is using <provided> for the spanner library, rather than including it themselves. You might need to put in a <excludes> and the earlier version of the spanner client. This doesn't mean that it's a drop in replacement, as our client library team has been know to change stuff.

[dzlier-gcp]

I have still not gotten spanner tests to pass, but since Vision is working I would like to submit that part so that test starts passing, and then continue to work on the spanner part.