[ACCUMULO-4053] ReplicationOperations.drain() is returning too quickly

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.7.1, 1.8.0
• Component/s: replication
• Assignee: Josh Elser
• Reporter: Josh Elser
• Created: 16/Nov/15 18:01
• Updated: 30/Dec/15 21:40
• Resolved: 30/Dec/15 21:40

Description

Looking into a failure of MultiInstanceReplicationIT.dataWasReplicatedToThePeer, the test recorded that the primary instance had data that the peer did not.

Digging into the logs, I can see that by the time the test had failed, the master was only closing and creating the work records for one of the WALs that the test should've been waiting on to drain.

As such, something is goofed up.