Hello -- I'm running a spark job which uses an HBase connection and running https://issues.apache.org/jira/browse/HBASE-19769 when using hbase-client version 1.4.0. The issue as I understand it has something to do with the way hadoop and / or spark are using multiple classloaders. I currently create a fat jar containing all the dependencies I need, including the hbase-client dependency and use this directly with spark-submit.
Can anyone suggest a workaround for this issue, or should I just switch back to 1.3.1? Thanks for your help, Morri Feldman AppsFlyer
