[HADOOP-13817] Add a finite shell command timeout to ShellBasedUnixGroupsMapping

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.6.0
• Fix Version/s: 2.9.0, 3.0.0-alpha4
• Component/s: security
• Hadoop Flags: Reviewed

Release Note

A new introduced configuration key "hadoop.security.groups.shell.command.timeout" allows applying a finite wait timeout over the 'id' commands launched by the ShellBasedUnixGroupsMapping plugin. Values specified can be in any valid time duration units: https://hadoop.apache.org/docs/current/api/org/apache/hadoop/conf/Configuration.html#getTimeDuration-java.lang.String-long-java.util.concurrent.TimeUnit-

Value defaults to 0, indicating infinite wait (preserving existing behaviour).

• Assignee: Harsh J
• Reporter: Harsh J
• Created: 14/Nov/16 11:49
• Updated: 02/Oct/19 17:13
• Resolved: 24/Feb/17 16:07

Description

The ShellBasedUnixGroupsMapping run various id commands via the ShellCommandExecutor modules without a timeout set (its set to 0, which implies infinite).

If this command hangs for a long time on the OS end due to an unresponsive groups backend or other reasons, it also blocks the handlers that use it on the NameNode (or other services that use this class). That inadvertently causes odd timeout troubles on the client end where its forced to retry (only to likely run into such hangs again with every attempt until at least one command returns).

It would be helpful to have a finite command timeout after which we may give up on the command and return the result equivalent of no groups found.

Issue Links

• breaks: HADOOP-15523 Shell command timeout given is in seconds whereas it is taken as millisec while scheduling
• links to: GitHub Pull Request #161