yeah I know. I can`t do nothing to improve my network, so, all i manage to do is: if it looks hanging, i kill it and restart. so far so good, but looks long way to go.
On Wed, Dec 18, 2013 at 3:35 PM, Azuryy Yu <azury...@gmail.com> wrote: > Hi Phoenix, > This is not Spark releated. It was your local net work limited. > > Thanks. > > > On Wed, Dec 18, 2013 at 3:17 PM, phoenix bai <mingzhi...@gmail.com> wrote: > >> I am compiling against hadoop 2.2.0, it really takes time, especially >> with network connection is not that stable and all. >> >> SPARK_HADOOP_VERSION=2.2.0 SPARK_YARN=true ./sbt/sbt assembly >> >> >> On Wed, Dec 18, 2013 at 10:39 AM, Patrick Wendell <pwend...@gmail.com>wrote: >> >>> Hey Philip, >>> >>> No - those are compiled against the "mr1" version. You'll need to >>> build yourself for YARN. >>> >>> - Patrick >>> >>> On Tue, Dec 17, 2013 at 10:32 AM, Philip Ogren <philip.og...@oracle.com> >>> wrote: >>> > I have a question about the pre-built binary for 0.8.0 for CDH 4 listed >>> > here: >>> > >>> > http://spark.incubator.apache.org/downloads.html >>> > >>> > and linked here: >>> > >>> > http://spark-project.org/download/spark-0.8.0-incubating-bin-cdh4.tgz >>> > >>> > Should I expect that this release is compiled with YARN enabled (i.e. >>> the >>> > SPARK_YARN property is set to true)? >>> > >>> > Thanks, >>> > Philip >>> >> >> >