[OOZIE-3527] Oozie stuck in waiting state if CoordPushDependencyCheckXCommand is not requeued

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 5.2.0
• Assignee: Manasa Gogineni
• Reporter: Rohini Palaniswamy
• Created: 11/Jul/19 16:56
• Updated: 20/Feb/20 13:02
• Resolved: 16/Jul/19 22:37

Description

We had a case where CoordPushDependencyCheckXCommand failed during loadState during a DB service failover to another node. Failure in loadState does not requeue. Ideally RecoveryService should pick it up, but since the CoordActionInputCheckXCommand was also running due to hdfs dependencies and constantly updating the lastModifiedTime, the RecoveryService was also not picking it up. Action was stuck in WAITING forever as the hcat dependencies were never discovered.