Henry Robinson b9034ea0d5 IMPALA-4580: Fix crash with FETCH_FIRST when #rows < result cache size ...

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).

This patch makes GetNext() a no-op if called after it sets *eos,
avoiding the crash..

Testing:
  Regression test that triggered the bug before this fix.

Change-Id: I454cd8a6cf438bdd0c49fd27c2725d8f6c43bb1d
Reviewed-on: http://gerrit.cloudera.org:8080/5335
Reviewed-by: Henry Robinson <henry@cloudera.com>
Reviewed-by: Dan Hecht <dhecht@cloudera.com>
Tested-by: Internal Jenkins

2016-12-03 11:07:04 +00:00

..

__init__.py

IMPALA-565: Support user impersonation for HS2 authorization requests

2014-01-08 10:53:39 -08:00

hs2_test_suite.py

IMPALA-3918: Remove Cloudera copyrights and add ASF license header

2016-08-09 08:19:41 +00:00

test_fetch_first.py

IMPALA-4580: Fix crash with FETCH_FIRST when #rows < result cache size

2016-12-03 11:07:04 +00:00

test_fetch.py

IMPALA-3918: Remove Cloudera copyrights and add ASF license header

2016-08-09 08:19:41 +00:00

test_hs2.py

IMPALA-4488: HS2 GetOperationStatus() should keep session alive

2016-11-24 05:39:45 +00:00

test_json_endpoints.py

IMPALA-2905: Handle coordinator fragment lifecycle like all others

2016-10-16 15:55:29 +00:00