[YARN-5773] RM recovery too slow due to LeafQueue#activateApplication()

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 2.8.0, 3.0.0-alpha2
• Component/s: capacity scheduler, rolling upgrade
• Labels: oct16-medium
• Hadoop Flags: Reviewed
• Assignee: Bibin Chundatt
• Reporter: Bibin Chundatt
• Created: 24/Oct/16 13:07
• Updated: 06/Jan/17 08:00
• Resolved: 30/Oct/16 06:22

Description

1. Submit application 10K application to default queue.
2. All applications are in accepted state
3. Now restart resourcemanager

For each application recovery LeafQueue#activateApplications() is invoked.Resulting in AM limit check to be done even before Node managers are getting registered.

Total iteration for N application is about N(N+1)/2 for 10K application 50000000 iterations causing time take for Rm to be active more than 10 min.

Since NM resources are not yet added to during recovery we should skip activateApplicaiton()

Issue Links

• breaks: YARN-5994 TestCapacityScheduler.testAMLimitUsage fails intermittently