yes, it is because REDUCE capability required is more than the supported max container capability in the cluster. Killing the Job. reduceResourceRequest: <memory:4096, vCores:1> maxContainerCapability:<memory:2048, vCores:8>. The problem is solved, when I change the memory to 4096 or a bigger number.
------------------ ???????? ------------------ ??????: "Li Yang";<liy...@apache.org>; ????????: 2016??5??15??(??????) ????7:35 ??????: "dev"<dev@kylin.apache.org>; ????: Re: killed by admin At least the hadoop job history server should have some traces. On Tue, May 10, 2016 at 11:13 AM, ???? <775620...@qq.com> wrote: > Hi all: > > When I build the cube, in the second step 'Extract Fact Table > Distinct Columns', the log shows 'killed by admin'. > And I could not find any error log.