you need to increase spark.yarn.executor.memoryOverhead. it has nothing to do with storage layer.
--Xuefu On Fri, Oct 23, 2015 at 4:49 AM, Jone Zhang <joyoungzh...@gmail.com> wrote: > I get an the error every time while I run a query on a large data set. I > think use MEMORY_AND_DISK can avoid this problem under the limited > resources. > "15/10/23 17:37:13 Reporter WARN > org.apache.spark.deploy.yarn.YarnAllocator>> Container killed by YARN for > exceeding memory limits. 7.6 GB of 7.5 GB physical memory used. Consider > boosting spark.yarn.executor.memoryOverhead." > > 2015-10-23 19:40 GMT+08:00 Xuefu Zhang <xzh...@cloudera.com>: > >> Yeah. for that, you cannot really cache anything through Hive on Spark. >> Could you detail more what you want to achieve? >> >> When needed, Hive on Spark uses memory+disk for storage level. >> >> On Fri, Oct 23, 2015 at 4:29 AM, Jone Zhang <joyoungzh...@gmail.com> >> wrote: >> >>> 1.But It's no way to set Storage Level through properties file in >>> spark, Spark provided "def persist(newLevel: StorageLevel)" >>> api only... >>> >>> 2015-10-23 19:03 GMT+08:00 Xuefu Zhang <xzh...@cloudera.com>: >>> >>>> quick answers: >>>> 1. you can pretty much set any spark configuration at hive using set >>>> command. >>>> 2. no. you have to make the call. >>>> >>>> >>>> >>>> On Thu, Oct 22, 2015 at 10:32 PM, Jone Zhang <joyoungzh...@gmail.com> >>>> wrote: >>>> >>>>> 1.How can i set Storage Level when i use Hive on Spark? >>>>> 2.Do Spark have any intention of dynamically determined Hive on >>>>> MapReduce or Hive on Spark, base on SQL features. >>>>> >>>>> Thanks in advance >>>>> Best regards >>>>> >>>> >>>> >>> >> >