AnsweredAssumed Answered

Impala 2.1 on MapR

Question asked by ananth on Dec 31, 2014
Latest reply on Oct 30, 2015 by jwenzel

I was wondering if there is any scheduled timeline for MapR to include Impala 2.1 in its releases ?

We are actively looking at Impala as our query mechanism into a hadoop based data warehouse.

We need the following features to implement our design :

 1. Impala Query execution engine with nested query support
 2. Impala Caching features coupled with HDFS caching
 3. Impala support for incremental compute stats
 4. HDFS Symbolic linking APIs enabled
 5. HDFS storage policies ( so that we can control the placement of our parquet files based on our predicted usage patterns - Ex: Recent data goes into SSDs and older data goes onto disk )

We have been looking at CDH for a few months and implemented quite a bit of code but hit a roadblock with respect to points 4 and 5 above.

I believe MapR would be supporting points 4 and 5 but I was disappointed to see Impala is still on 1.4 versions for the latest MapR distributions ( Atleast the VM points to an older release of Impala).

I was wondering if there is a roadmap for impala upgrades inside new MapR distributions ?

Any pointers for expected timelines would immensely help us in planning our hadoop infrastructure.

Thanks for your time.