[HDFS-7638] Small fix and few refinements for FSN#truncate

• Type: Sub-task
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.7.0
• Component/s: datanode, namenode
• Hadoop Flags: Reviewed
• Assignee: Yi Liu
• Reporter: Yi Liu
• Created: 19/Jan/15 08:20
• Updated: 24/Apr/15 23:22
• Resolved: 20/Jan/15 01:40

Description

1.

removeBlocks(collectedBlocks);

should be after logSync, as we do in other FSN places (rename, delete, write with overwrite), the reason is discussed in HDFS-2815 and https://issues.apache.org/jira/browse/HDFS-6871?focusedCommentId=14110068&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14110068

2.

stat = FSDirStatAndListingOp.getFileInfo(dir, src, false,
    FSDirectory.isReservedRawName(src), true);

We'd better to use dir.getAuditFileInfo, since it's only for audit log. If audit log is not on, we don't need to get the file info.

3.
In truncateInternal,

INodeFile file = iip.getLastINode().asFile();

is not necessary.