Hi, Do you have the same problem on SPARK_HOME/bin/spark-shell? Are you using standalone spark cluster? or Yarn?
Thanks, moon On Sun, Nov 13, 2016 at 8:19 PM York Huang <yorkhuang.d...@gmail.com> wrote: > I ran into the "No space left on device" error in zeppelin spark when I > tried to run the following. > cache table temp_tbl as > select * from > ( > select *, rank() over (partition by id order by year desc) as rank > from table1 > ) v > where v.rank =1 > > The table1 is very big. > > I set up spark.local.dir in spark-env.sh. > SPARK_LOCAL_DIRS in spark-defaults.conf, > and spark.local.dir in zeppelin spark interpreter parameter. They are all > set to /apps/spark_tmp folder. > > I can see some folders are created in /apps/spark_tmp. But I can still see > in df -h, /tmp is running out of space. > > What is missing? > > Thanks >