Sure that can be configurable. Kickoff the work by open a JIRA. Thanks Yang
On Tue, Apr 17, 2018 at 10:21 AM, <[email protected]> wrote: > Currently in Spark cubing, the StorageLevel is set to > StorageLevel.MEMORY_AND_DISK_SER, which will take up a lot of memory if > the RDD of the layer is large. Can we make StorageLevel configurable ? So > that for large cube, user can set it to Disk to avoid OOM error. >
