[ 
https://issues.apache.org/jira/browse/HBASE-20188?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16398100#comment-16398100
 ] 

Appy commented on HBASE-20188:
------------------------------

I was planning to do the same - test with YCSB.

How does following plan sounds? 1Master + 6 RS + 2 client.

Begin with hbase1.4, load 1B rows (yscb load phase) , flush, compact, take 
snapshot.

run workload : write only (custom), C, E, F ([list of 
workloads|https://github.com/brianfrankcooper/YCSB/wiki/Core-Workloads])

Update to 2.0, run same workloads and compare.

Do one off 1B load on 2.0 to compare with initial load on 1.4.

 

What remains:
 * perf for increments
 * these will be done with basic configs (maybe I'll just increase heap size), 
so it won't test fancy stuff - off heap, etc.
 * what else?

  

> [TESTING] Performance
> ---------------------
>
>                 Key: HBASE-20188
>                 URL: https://issues.apache.org/jira/browse/HBASE-20188
>             Project: HBase
>          Issue Type: Umbrella
>          Components: Performance
>            Reporter: stack
>            Priority: Critical
>             Fix For: 2.0.0
>
>
> How does 2.0.0 compare to old versions? Is it faster, slower? There is rumor 
> that it is much slower, that the problem is the asyncwal writing. Does 
> in-memory compaction slow us down or speed us up? What happens when you 
> enable offheaping?
> Keep notes here in this umbrella issue. Need to be able to say something 
> about perf when 2.0.0 ships.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to