python - Installing pyspark on hadoop and yarn -
i have installed spark on top of hadoop , yarn. when launch pyspark shell , try compute error.
error python worker: /usr/bin/python: no module named pyspark the pyspark module should there, have put external link it?
i have encountered same problem. due building spark java7. using java6 should fix it. it's python's fault - zipimport module not support zip64 format used java7 package large jar 65536+ files. google hit helped me. http://mail-archives.apache.org/mod_mbox/incubator-spark-user/201406.mbox/%3ccamjob8kcgk0pqiogju6uokceyswcusw3xwd5wrs8ikpmgd2dag@mail.gmail.com%3e
this explains it. https://issues.apache.org/jira/browse/spark-1911
Comments
Post a Comment