Can you pastebin the class path ? Thanks
> On May 24, 2015, at 5:02 AM, boci <boci.b...@gmail.com> wrote: > > Yeah, I have same jar with same result, I run in docker container and I using > same docker container with my another project... the only difference is the > postgresql jdbc driver and the custom RDD... no additional dependencies (both > single jar generated with same assembly configuration with same dependencies) > and the second is work like a charm > > Another idea? > > 2015. máj. 24. 2:41 ezt írta ("Ted Yu" <yuzhih...@gmail.com>): >> In my local maven repo, I found: >> >> $ jar tvf >> /Users/tyu/.m2/repository//org/spark-project/akka/akka-actor_2.10/2.3.4-spark/akka-actor_2.10-2.3.4-spark.jar >> | grep SelectionPath >> 521 Mon Sep 29 12:05:36 PDT 2014 akka/actor/SelectionPathElement.class >> >> Is the above jar in your classpath ? >> >>> On Sat, May 23, 2015 at 5:05 PM, boci <boci.b...@gmail.com> wrote: >>> Hi guys! >>> >>> I have a small spark application. It's query some data from postgres, >>> enrich it and write to elasticsearch. When I deployed into spark container >>> I got a very fustrating error: >>> https://gist.github.com/b0c1/66527e00bada1e4c0dc3 >>> >>> Spark version: 1.3.1 >>> Hadoop version: 2.6.0 >>> Additional info: >>> serialization: kryo >>> rdd: custom rdd to query >>> >>> I not understand >>> 1. akka.actor.SelectionPath doesn't exists in 1.3.1 >>> 2. I checked all dependencies in my project, I only have >>> "org.spark-project.akka:akka-*_2.10:2.3.4-spark:jar" doesn't have >>> 3. I not found any reference for this... >>> 4. I created own RDD, it's work, but I need to register to kryo? (mapRow >>> using ResultSet, I need to create >>> 5. I used some months ago and it's already worked with spark 1.2... I >>> recompiled with 1.3.1 but I got this strange error >>> >>> Any idea? >>> >>> ---------------------------------------------------------------------------------------------------------------------------------- >>> Skype: boci13, Hangout: boci.b...@gmail.com