May be your one reducer is overloaded due to groupby keys. If you are using groupby then try below property and see if reducer data distributed.
set hive.groupby.skewindata=true; Thanks Jitendra On Mon, May 11, 2015 at 12:35 PM, r7raul1...@163.com <r7raul1...@163.com> wrote: > Status: Running (Executing on YARN cluster with App id > application_1419300485749_1493279) > > -------------------------------------------------------------------------------- > > VERTICES STATUS TOTAL COMPLETED RUNNING PENDING FAILED KILLED > -------------------------------------------------------------------------------- > > Map 1 .......... SUCCEEDED 200 200 0 0 0 0 > Map 4 .......... SUCCEEDED 3 3 0 0 0 0 > Map 5 .......... SUCCEEDED 152 152 0 0 0 0 > Reducer 2 ..... RUNNING 20 19 1 0 0 0 > Reducer 3 RUNNING 23 0 23 0 0 0 > -------------------------------------------------------------------------------- > > VERTICES: 03/05 [========================>>--] 93% ELAPSED TIME: 791.14 s > > > A reduce run for a long time. > > I try set hive.exec.reducers.bytes.per.reducer = 4000000000; > set hive.skewjoin.key = 1000000000; > set hive.optimize.skewjoin =true; > > but nothing helped. Only the reduce num decrease.... > > > ------------------------------ > r7raul1...@163.com >
