[HADOOP-13442] Optimize UGI group lookups

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.8.0, 3.0.0-alpha1, 2.7.5
• Target Version/s: 2.9.0
• Hadoop Flags: Reviewed
• Assignee: Daryn Sharp
• Reporter: Daryn Sharp
• Created: 29/Jul/16 17:35
• Updated: 21/Jun/20 19:36
• Resolved: 04/Aug/16 15:50

Description

UGI#getGroups and its usage is inefficient. The list is unnecessarily converted to multiple collections.

For every invocation, the List<String> from the group provider is converted into a LinkedHashSet<String> (to de-dup), back to a String[]. Then callers testing for group membership convert back to a List<String>. This should be done once to reduce allocations.

Issue Links

• breaks: HDFS-10738 Fix TestRefreshUserMappings.testRefreshSuperUserGroupsConfiguration test failure
• breaks: MAPREDUCE-6750 TestHSAdminServer.testRefreshSuperUserGroups is failing
• is duplicated by: HADOOP-12253 ViewFileSystem getFileStatus java.lang.ArrayIndexOutOfBoundsException: 0
• is duplicated by: HADOOP-12858 Reduce UGI getGroups overhead
• is required by: HDFS-10711 Optimize FSPermissionChecker group membership check
• relates to: HADOOP-17079 Optimize UGI#getGroups by adding UGI#getGroupsSet