AnsweredAssumed Answered

Does MapR support Hadoop 2.3.0 solution for in-memory data at datanodes, i.e.HDFS-4949?

Question asked by vikasn on Mar 11, 2014
Latest reply on Apr 8, 2014 by Ted Dunning
Hi,

Reposting this question again - ( unable to locate previous post).

Wonder if MapR has plans to allow specific file paths to be cached in memory on the MapR cluster side. This is a useful case, and well illustrated in HDFS-4949 issue, it's cache requirement document.

https://issues.apache.org/jira/i#browse/HDFS-4949

One possibility is allow a volume to be cached and pinned to memory. I wonder if Spark may help but that needs another layer (other packages - hive) for smaller use cases.

Thanks.

Outcomes