hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "vincent eiger (JIRA)" <j...@apache.org>
Subject [jira] [Created] (HBASE-17340) Workload management with Hbase
Date Mon, 19 Dec 2016 15:16:59 GMT
vincent eiger created HBASE-17340:

             Summary: Workload management with Hbase 
                 Key: HBASE-17340
                 URL: https://issues.apache.org/jira/browse/HBASE-17340
             Project: HBase
          Issue Type: Brainstorming
          Components: Client
            Reporter: vincent eiger

Hi guys, I am project manager in a bank where we work on a POC. We try to test the first step
of our Datalake with:
- a first layer for landing area : store files without transformation in HDFS (nowdays 6 files)
-a second layer in Hbase where we upload with one spark job a unique Hbase table.
We know that with capacity scheduler in Yarn we could guarantee resources consumption (memory,
CPU) for spark jobs.
My question is :
- how i could guarantee resources for my Hbase client : I mean one application which request
my table with Hbase native client and at the same time some other application, batch run spark
jobs... I want ensure all the time at least 30% of capactity to my application (which request
my Hbase table) in order to guarantee for instance 50 request at the same time with less than
200 ms per user.
Is Slider the good answer ?
Thanks a lot for answers, 

This message was sent by Atlassian JIRA

View raw message