[AURORA-740] Be more graceful when failing over

• Type: Story
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: Scheduler
• Assignee: Unassigned
• Reporter: Bill Farner
• Created: 23/Sep/14 00:27
• Updated: 03/Jul/15 07:46

Description

The scheduler fails over periodically, which can cause timed out tasks when scheduler->master messages are still in the local queue while the scheduler tears down. We could probably mitigate this by tearing down the scheduler in a way such that we allow this queue some time to drain before stopping the driver.