[FLUME-1320] Add safeguard for checkpoint corruption detection

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.2.0
• Assignee: Arvind Prabhakar
• Reporter: Arvind Prabhakar
• Created: 26/Jun/12 07:30
• Updated: 26/Jun/12 11:03
• Resolved: 26/Jun/12 09:25

Description

The checkpoint process can potentially be interrupted by a catastrophic failure such as complete power loss leaving the checkpoint corrupted. When the channel is restarted, the (corrupted) checkpoint will get loaded and there could be dataloss as a result of this.

It is therefore necessary to have a safeguard to detect if a checkpoint succeeded or failed.