[HDFS-15882] Fix incorrectly initializing RandomAccessFile based on configuration options

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.4.0
• Fix Version/s: 3.4.0
• Component/s: namenode
• Labels: pull-request-available
• Target Version/s: 3.4.0
• Hadoop Flags: Reviewed
• Assignee: Xie Lei
• Reporter: Xie Lei
• Created: 07/Mar/21 07:08
• Updated: 16/Oct/24 18:29
• Resolved: 13/Mar/21 06:24

Description

`rw` Open for reading and writing. If the file does not already exist then an attempt will be made to create it.
`rws` Require that every update to the file's content or metadata be written synchronously to the underlying storage device.
From the literal meaning of this variable `shouldSyncWritesAndSkipFsync`, we should use `rw` when `shouldSyncWritesAndSkipFsync` is true.

We use SATA disk to store the journal node's data. It's not effective for improving RPC performance whether the `shouldSyncWritesAndSkipFsync` variable is true or false. it's caused by initializing RandomAccessFile incorrectly.

Issue Links

• causes: HDFS-16910 Fix incorrectly initializing RandomAccessFile caused flush performance decreased for JN
• duplicates: HDFS-15881 Fix incorrectly initializing RandomAccessFile based on configuration options
• links to: GitHub Pull Request #2751