[IMPALA-5691] test_low_mem_limit_q18 is flaky

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: Impala 2.9.0
• Fix Version/s: Impala 2.10.0
• Labels: flaky
• Epic Color: ghx-label-5
• Assignee: Tim Armstrong
• Reporter: Anuj Phadke
• Created: 21/Jul/17 08:12
• Updated: 22/May/18 18:28
• Resolved: 26/Jul/17 22:43

Description

test_low_mem_limit_q18 failed while running this job -
https://jenkins.impala.io/job/gerrit-verify-dryrun/900/console

This does not consistently happen though.

22:55:47 ] TestTpchMemLimitError.test_low_mem_limit_q18[mem_limit: 275 | exec_option:

{'batch_size': 0, 'num_nodes': 0, 'disable_codegen_rows_threshold': 0, 'disable_codegen': False, 'abort_on_error': 1, 'exec_single_node_rows_threshold': 0}

| table_format: parquet/none]
22:55:47 ] [gw1] linux2 – Python 2.7.6 /home/ubuntu/Impala/bin/../infra/python/env/bin/python
22:55:47 ] query_test/test_mem_usage_scaling.py:187: in test_low_mem_limit_q18
22:55:47 ] self.low_memory_limit_test(vector, 'tpch-q18', self.MIN_MEM_FOR_TPCH['Q18'])
22:55:47 ] query_test/test_mem_usage_scaling.py:101: in low_memory_limit_test
22:55:47 ] assert TestLowMemoryLimits.EXPECTED_ERROR_MSG in str(e)
22:55:47 ] E assert 'Memory limit exceeded' in 'ImpalaBeeswaxException:\n Query aborted:Cannot perform aggregation at node with id 14. Repartitioning did not reduce the size of a spilled partition. Repartitioning level 6. Number of rows 1.\n\n'
22:55:47 ] E + where 'Memory limit exceeded' = TestLowMemoryLimits.EXPECTED_ERROR_MSG
22:55:47 ] E + and 'ImpalaBeeswaxException:\n Query aborted:Cannot perform aggregation at node with id 14. Repartitioning did not reduce the size of a spilled partition. Repartitioning level 6. Number of rows 1.\n\n' = str(ImpalaBeeswaxException())
22:55:47 ] ----------------------------- Captured stderr call -----------------------------
22:55:47 ] – executing against localhost:21000
22:55:47 ] use tpch_parquet;
22:55:47 ]
22:55:47 ] SET batch_size=0;
22:55:47 ] SET num_nodes=0;
22:55:47 ] SET disable_codegen_rows_threshold=0;
22:55:47 ] SET disable_codegen=False;
22:55:47 ] SET abort_on_error=1;
22:55:47 ] SET mem_limit=275m;
22:55:47 ] SET exec_single_node_rows_threshold=0;
22:55:47 ] MainThread: Query Name:
22:55:47 ] TPCH-Q18
22:55:47 ]
22:55:47 ] – executing against localhost:21000
22:55:47 ] select
22:55:47 ] c_name,
22:55:47 ] c_custkey,
22:55:47 ] o_orderkey,
22:55:47 ] o_orderdate,
22:55:47 ] o_totalprice,
22:55:47 ] sum(l_quantity)
22:55:47 ] from
22:55:47 ] customer,
22:55:47 ] orders,
22:55:47 ] lineitem
22:55:47 ] where
22:55:47 ] o_orderkey in (
22:55:47 ] select
22:55:47 ] l_orderkey
22:55:47 ] from
22:55:47 ] lineitem
22:55:47 ] group by
22:55:47 ] l_orderkey
22:55:47 ] having
22:55:47 ] sum(l_quantity) > 300
22:55:47 ] )
22:55:47 ] and c_custkey = o_custkey
22:55:47 ] and o_orderkey = l_orderkey
22:55:47 ] group by
22:55:47 ] c_name,
22:55:47 ] c_custkey,
22:55:47 ] o_orderkey,
22:55:47 ] o_orderdate,
22:55:47 ] o_totalprice
22:55:47 ] order by
22:55:47 ] o_totalprice desc,
22:55:47 ] o_orderdate
22:55:47 ] limit 100;
22:55:47 ]
22:55:47 ] ======= 1 failed, 1436 passed, 42 skipped, 36 xfailed in 1806.79 seconds =======

Issue Links

• relates to: IMPALA-2708 Partitioned aggregation node repartitions when spilled partition could fit in memory