Thanks for your reply. Nope, nothing has changed, i mean we didn't change any kylin configuration recently. We are using snappy compression for hbase tables "kylin.hbase.default.compression.codec=snappy" and this has been there for almost 3 years. All the other cubes are running fine with same configs. This cube has been having issues with build dimension dictionary step, there are a couple of measures of UHC columns on this cube with ~40 million records. To make the cube build, i bumped up the kylin job node from 64 GB to 160 GB node and the build dimension step is running fine now but the cube is failing at "#12 Step Name: Build N-Dimension Cuboid Data : 22-Dimension" Also another weird thing is the cube has been building fine until last week, the data increased by only about 200K records. Any help is greatly appreciated.
-- Sent from: http://apache-kylin.74782.x6.nabble.com/
