Fixed Issues in Apache Hadoop

Review the list of Hadoop issues that are resolved in Cloudera Runtime 7.1.7 SP1.

CDPD-34911: Harden FileUtil.unTar(File inFile, File untarDir).

This issue is now resolved.

CDPD-27264: S3A Fileystem creation fails "Unable to find a region via the region provider chain."

This issue is now resolved.

CDPD-26975: Configuration ${env.VAR:-FALLBACK} should eval FALLBACK when restrictSystemProps=true.

This issue is now resolved.

CDPD-26548: Intermittent S3AInputStream failures: Premature end of Content-Length.

This issue is now resolved.

CDPD-26441: This introduces "fs.s3a.endpoint.region", which can be used to set the region of the S3 bucket, and bypasses the parsing of the region from an endpoint.

This issue is now resolved.

CDPD-21205: Collect AwsSdkMetrics in S3A FileSystem IOStatistics.

This issue is now resolved.

CDPD-18287: Reduced thread contention while initialising filesystems, especially AzureBlobFileSystem or other object stores; The number of filesystems which can be created in parallelis now limited in the option "fs.creation.parallel.count" - default value 64. A smaller value may improve worker thread startup time in processes which have many worker threads all attempting to interact with the same object store(s)

This issue is now resolved.

CDPD-13652: HADOOP-17195. The abfs connector's output streams can buffer blocks to write on disk (default), byte array or off-heap bytebuffer. Configuration options allow queue length and size of shared pool of uploader threads to set

This issue is now resolved.