[YARN-8581] [AMRMProxy] Add sub-cluster timeout in LocalityMulticastAMRMProxyPolicy

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.10.0, 3.2.0
• Component/s: amrmproxy, federation
• Assignee: Botong Huang
• Reporter: Botong Huang
• Created: 25/Jul/18 22:35
• Updated: 15/Oct/18 13:42
• Resolved: 21/Aug/18 20:07

Description

In Federation, every time an AM heartbeat comes in, LocalityMulticastAMRMProxyPolicy in AMRMProxy splits the asks according to the list of active and enabled sub-clusters. However, if we haven't been able to heartbeat to a sub-cluster for some time (network issues, or we keep hitting some exception from YarnRM, or YarnRM master-slave switch is taking a long time etc.), we should consider the sub-cluster as unhealthy and stop routing asks there, until the heartbeat channel becomes healthy again.