[HELIX-350] cluster status monitor should not be reset in FINALIZE type pipeline

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 0.7.1, 0.6.4
• Assignee: Shi Lu
• Reporter: Shi Lu
• Created: 09/Jan/14 22:34
• Updated: 25/Aug/14 17:37
• Resolved: 05/Apr/14 17:49

Description

Currently, the cluster status monitor is disposed when the controller is running a FINALIZE type pipeline, which is possible when a participant is down (which make the controller no-longer listen to its current state change and will trigger FINALIZE type pipeline)

the change is to dispose the cluster status monitor when the controller helix manager disconnects from zookeeper.