Hey Guys, do you know I can run Presto on MapR cluster? I was thinking to play with it, but just wanted to find out if there is anybody else who successfully running Presto?
I don't think you can do it, but I really hope MapR gets this integration done soon. I hope to see an answer that I'm wrong though.
AFAIK, presto runs off cluster on its own cluster. The issue would be that it expects HDFS and not MapRFS.
Of course I don't play with presto... YMMV and I'm just guessing. ;-)
Wow, thats not good Thanks for reply anyway.
What are you trying to do? MapR typically defers to Drill if you are looking to do SQL queries on the platform.
I want to have a sql query engine with ODBC support for running ad-hoc queries. I am playing now with Drill, but wanted to do some performance comparison with presto.
Have you seen this article?
How does Apache Drill compare to Facebook's Presto? - Quora
I work for a company called Kognitio and we provide a SQL engine that runs on MapR and connects via ODBC or JDBC.
We use in-memory computing techniques to provide very high performance as well as other great features such as a mature ANSI SQL implementation and SQL extensions to allow integrated data science analysis. Due to the in-memory architecture, it doesn't require indexes which makes it ideal for ad-hoc use.
Thanks Mark, will have a look at the website.
Shameful plug? ;-P
You can run Kognitio on the MapR cluster or off.
If you run it on cluster, you need to account for the memory allocation so that you don't have an over subscription of resources.
And if you run out of memory, Kognitio should spill to disk.
Also... w Kognitio, size matters. Tomasz never said how much data...
Retrieving data ...