[HADOOP-12444] Support lazy seek in S3AInputStream

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.7.1
• Fix Version/s: 2.8.0, 3.0.0-alpha1
• Component/s: fs/s3
• Assignee: Rajesh Balamohan
• Reporter: Rajesh Balamohan
• Created: 28/Sep/15 00:54
• Updated: 17/Apr/19 19:12
• Resolved: 09/Apr/16 10:26

Description

• Currently, "read(long position, byte[] buffer, int offset, int length)" is not implemented in S3AInputStream (unlike DFSInputStream). So, "readFully(long position, byte[] buffer, int offset, int length)" in S3AInputStream goes through the default implementation of seek(), read(), seek() in FSInputStream.
• However, seek() in S3AInputStream involves re-opening of connection to S3 everytime (https://github.com/apache/hadoop/blob/trunk/hadoop-tools/hadoop-aws/src/main/java/org/apache/hadoop/fs/s3a/S3AInputStream.java#L115).
• It would be good to consider having a lazy seek implementation to reduce connection overheads to S3. (e.g Presto implements lazy seek. https://github.com/facebook/presto/blob/master/presto-hive/src/main/java/com/facebook/presto/hive/PrestoS3FileSystem.java#L623)

Issue Links

• contains: HADOOP-11874 s3a can throw spurious IOEs on close()
• contains: HADOOP-12976 s3a toString to be meaningful in logs
• depends upon: HADOOP-12994 Specify PositionedReadable, add contract tests, fix problems
• is related to: HADOOP-16241 S3AInputStream PositionReadable should perform ranged read on dedicated stream
• is required by: HADOOP-12844 Recover when S3A fails on IOException in read()
• relates to: HIVE-11945 ORC with non-local reads may not be reusing connection to DN