[HDFS-1028] INode.getPathNames could split more efficiently

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 0.22.0
• Component/s: namenode
• Hadoop Flags: Reviewed
• Assignee: Dmytro Molkov
• Reporter: Todd Lipcon
• Created: 06/Mar/10 07:06
• Updated: 07/May/10 05:34
• Resolved: 07/May/10 05:34

Description

INode.getPathnames uses String.split(String) which actually uses the full Java regex implementation. Since we're always splitting on a single char, we could implement a faster one like StringUtils.split() (except without the escape character). This takes a significant amount of CPU during FSImage loading so should be a worthwhile speedup.

Issue Links

• is blocked by: HADOOP-6623 Add StringUtils.split for non-escaped single-character separator