[IMPALA-4580] Crash when FETCH_FIRST exhausts result set cache

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: Impala 2.8.0
• Fix Version/s: Impala 2.8.0
• Component/s: Backend
• Target Version: Impala 2.8.0
• Assignee: Henry Robinson
• Reporter: Henry Robinson
• Created: 02/Dec/16 23:27
• Updated: 05/Dec/16 05:14
• Resolved: 05/Dec/16 05:14

Description

The following sequence can lead to a crash:

1. Client sets result cache size to N
2. Client issues query with #results < N
3. Client fetches all results, triggering eos and tearing down Coordinator::root_sink_.
4. Client restarts query with FETCH_FIRST.
5. Client reads all results again. After cache is exhausted, Coordinator::GetNext() is called to detect eos condition again.
6. GetNext() hits DCHECK(root_sink_ != nullptr).

Either the QES or the coordinator need to remember that eos was once true, and turn GetNext() into a no-op if so.