[ https://issues.apache.org/jira/browse/SPARK-12453?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15067115#comment-15067115 ]
Sean Owen commented on SPARK-12453: ----------------------------------- OK, I see what happened here: https://github.com/apache/spark/commit/87f82a5fb9c4350a97c761411069245f07aad46f How about updating to 1.9.40 for consistency? really, it sounds like there's no point manually setting the SDK version here -- how about preemptively bringing those parts of SPARK-12269 back? Then really it should go into master first, and be backported, and then further updated by 12269. This is why I view it as sort of a duplicate, since it could as well come from back-porting just a subset of 12269. I don't know if a new 1.5.x release will happen. > Spark Streaming Kinesis Example broken due to wrong AWS Java SDK version > ------------------------------------------------------------------------ > > Key: SPARK-12453 > URL: https://issues.apache.org/jira/browse/SPARK-12453 > Project: Spark > Issue Type: Bug > Components: Streaming > Affects Versions: 1.5.2 > Reporter: Martin Schade > Priority: Critical > Labels: easyfix > > The Spark Streaming Kinesis Example (kinesis-asl) is broken due to wrong AWS > Java SDK version (1.9.16) referenced with the AWS KCL version (1.3.0). > AWS KCL 1.3.0 references AWS Java SDK version 1.9.37. > Using 1.9.16 in combination with 1.3.0 does fail to get data out of the > stream. > I tested Spark Streaming with 1.9.37 and it works fine. > Testing a simple KCL client outside of Spark with 1.3.0 and 1.9.16 also > fails, so it is due to the specific versions used in 1.5.2 and not a Spark > related implementation. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org