Hi, in this pull request Spark adds the ability to store RDDs in off heap memory via Tachyon: https://github.com/apache/incubator-spark/pull/468
I was wondering if the broadcast variable API would be opened up in a similar manner? From the source of e.g. HttpBroadcast, it is currently hard coded which would prevent Tachyon storage: SparkEnv.get.blockManager.putSingle(blockId, value_, StorageLevel.MEMORY_AND_DISK, false) Thanks, Adam -- View this message in context: http://apache-spark-user-list.1001560.n3.nabble.com/Tachyon-RDD-Caching-tp1820.html Sent from the Apache Spark User List mailing list archive at Nabble.com.
