XmlIOIT fails on Spark and Flink

When run on Spark or Flink remote cluster, XmlIOIT fails with the following stacktrace:

 
```

org.apache.beam.sdk.io.xml.XmlIOIT > writeThenReadAll FAILED
java.lang.AssertionError: Calculate hashcode/Flatten.PCollections.out:
Expected:
"7f51adaf701441ee83459a3f705c1b86"
but: was "f45a386eeab2ea5c4a66c6aae07c88ae"
at org.apache.beam.sdk.testing.PAssert$PAssertionSite.capture(PAssert.java:168)
at
org.apache.beam.sdk.testing.PAssert.thatSingleton(PAssert.java:413)
at org.apache.beam.sdk.testing.PAssert.thatSingleton(PAssert.java:404)
at
org.apache.beam.sdk.io.xml.XmlIOIT.writeThenReadAll(XmlIOIT.java:140)

Caused by:
java.lang.AssertionError:
Expected:
"7f51adaf701441ee83459a3f705c1b86"
but: was "f45a386eeab2ea5c4a66c6aae07c88ae"
```

Imported from Jira [BEAM-5165](https://issues.apache.org/jira/browse/BEAM-5165). Original Jira may contain additional context.
Reported by: ŁukaszG.