parquet-hadoop tests fail if working behind a web proxy

[alexeyroytman]

Describe the enhancement requested

My working environment is behind a web proxy. The maven build for me fails on parquet-hadoop tests that use files from https://github.com which is not directly accessible to me.

The failed tests are:

[ERROR]   TestInterOpReadByteStreamSplit.testReadAllSupportedTypes » SocketTimeout Connect timed out
[ERROR]   TestInterOpReadByteStreamSplit.testReadFloats » SocketTimeout Connect timed out
[ERROR]   TestInterOpReadFloat16.testInterOpReadFloat16NonZerosAndNansParquetFiles » SocketTimeout Connect timed out
[ERROR]   TestInterOpReadFloat16.testInterOpReadFloat16ZerosAndNansParquetFiles » SocketTimeout Connect timed out
[ERROR]   TestInteropBloomFilter.testReadDataIndexBloomParquetFiles » SocketTimeout Connect timed out
[ERROR]   TestInteropBloomFilter.testReadDataIndexBloomWithLengthParquetFiles » SocketTimeout Connect timed out
[ERROR]   TestInteropReadLz4RawCodec.testInteropReadLz4RawLargerParquetFiles » SocketTimeout Connect timed out
[ERROR]   TestInteropReadLz4RawCodec.testInteropReadLz4RawSimpleParquetFiles » SocketTimeout Connect timed out

They all are similar:

org.apache.parquet.hadoop.TestInterOpReadByteStreamSplit.testReadAllSupportedTypes -- Time elapsed: 10.24 s <<< ERROR!
java.net.SocketTimeoutException: Connect timed out
        at java.base/sun.nio.ch.NioSocketImpl.timedFinishConnect(NioSocketImpl.java:551)
        at java.base/sun.nio.ch.NioSocketImpl.connect(NioSocketImpl.java:602)
        at java.base/java.net.SocksSocketImpl.connect(SocksSocketImpl.java:327)
        at java.base/java.net.Socket.connect(Socket.java:633)
        at okhttp3.internal.platform.Platform.connectSocket(Platform.kt:128)
        at okhttp3.internal.connection.RealConnection.connectSocket(RealConnection.kt:295)
        at okhttp3.internal.connection.RealConnection.connect(RealConnection.kt:207)
        at okhttp3.internal.connection.ExchangeFinder.findConnection(ExchangeFinder.kt:226)

The main reason for this is that OkHttpClient for these tests is not configured to use any proxy, and it does not recognize system properties (e.g. https.proxyHost) or environment variables (e.g. https_proxy) for such flow.

Component(s)

Build

[alexeyroytman]

I'm going to fix it and to file a PR.

[alexeyroytman]

By the way, MAVEN_OPTS="-Dhttps.proxyHost=... -Dhttps.proxyPort=..." are not passed to the parquet-hadoop tests, thus I just append these definitions to the ./mvnw command line.

[alexeyroytman]

PR #3483