[ASTERIXDB-2112] Avoid aborting jobs, temporary UNUSABLE cluster on reestablished NC->CC IPC connection where no reboot took place

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: *DB - AsterixDB, CLUS - Cluster management
• Assignee: Unassigned
• Reporter: Michael Blow
• Created: 28/Sep/17 15:35
• Updated: 28/Sep/17 15:35

Description

See attached logs- when nc1's connection was reset, all jobs on that were running on that nc were aborted, and the cluster was temporarily UNUSABLE. Both of these should be able to be avoided in the case of re-registration without restart.