Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
Impala 2.7.0
Description
I've noticed that, starting from 2.7.0, the RowsProduced counter of the coordinator fragment always show 0, regardless of the number of rows returned by the query. For example, running the query "select * from sample_07" I get the below.
The full profile files are attached.
Coordinator Fragment F01 - AverageThreadTokens: 0.0 - BloomFilterBytes: 0 - InactiveTotalTime: 0 - PeakMemoryUsage: 88592 - PerHostPeakMemUsage: 0 - PrepareTime: 0 - RowsProduced: 0 - TotalCpuTime: 95999318 - TotalNetworkReceiveTime: 108999227 - TotalNetworkSendTime: 0 - TotalStorageWaitTime: 0 - TotalTime: 108999227 ... EXCHANGE_NODE (id=1) - BytesReceived: 40557 - ConvertRowBatchTime: 0 - DeserializeRowBatchTimer: 0 - FirstBatchArrivalWaitTime: 108999227 - InactiveTotalTime: 108999227 - PeakMemoryUsage: 0 - RowsReturned: 823 - RowsReturnedRate: 7550 - SendersBlockedTimer: 0 - SendersBlockedTotalTimer(*): 0 - TotalTime: 108999227
This seems to have started on 2.7.0