hive-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Larson, Kurt" <klar...@wbgames.com>
Subject RE: EMR 5.8 & Hue/Hive Performance/Stability Specifics
Date Tue, 12 Sep 2017 19:10:06 GMT
EMR switched its default execution engine from MR to Tez in EMR 5.0.0.  The current EMR is
5.8.0 and it still used Tez by default.

From: Jörn Franke [mailto:jornfranke@gmail.com]
Sent: Tuesday, September 12, 2017 12:47 PM
To: user@hive.apache.org
Subject: Re: EMR 5.8 & Hue/Hive Performance/Stability Specifics

Test it, because it really depends what you do. Since you use hue you seem to be interested
in interactive analysis, so the best is to use Tez and llap as a hive engine. Make also sure
that you use ORC or Parquet as a Hive storage format. Leverage the in-build orc or parquet
indexes by sorting data on the filtering column. If possible partition the data.

I do not remember  at the moment if TEZ is included in the Hive package of EMR or if you have
to install an additional package.
However, given your current Hive version you should see for sure an improvement.

On 12. Sep 2017, at 18:41, Mr. Glenn Waldman <glennwaldman@gmail.com<mailto:glennwaldman@gmail.com>>
wrote:
Hi,
We are upgrading our EMR from 4.7 - 5.8, along with Hue/Hive conjointly.
Hue: from 1.0 - 2.3
Hive: to 3.12

Can somebody explain some high-level, specific stability and performance increases we should
see from these upgrades?

We run daily queries through Hue/Hive and are looking for how these upgrades should make them
run more quickly with less breaks.

Thoughts!?

Thanks!!
Mime
View raw message