You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Ilya Suntsov (JIRA)" <ji...@apache.org> on 2015/12/04 12:21:11 UTC
[jira] [Commented] (IGNITE-2084) Yardstick: benchmark-run-all.sh
script should able to deploy and grab logs for distributed environment
[ https://issues.apache.org/jira/browse/IGNITE-2084?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15041437#comment-15041437 ]
Ilya Suntsov commented on IGNITE-2084:
--------------------------------------
Also would be nice if benchmark-run-all.sh be able to:
- start dstat on yardstick's servers
- calculate average (op/s) for each benchmark and print it in terminal after run
> Yardstick: benchmark-run-all.sh script should able to deploy and grab logs for distributed environment
> ------------------------------------------------------------------------------------------------------
>
> Key: IGNITE-2084
> URL: https://issues.apache.org/jira/browse/IGNITE-2084
> Project: Ignite
> Issue Type: Task
> Affects Versions: ignite-1.4, 1.5
> Reporter: Sergey Kozlov
> Assignee: Yakov Zhdanov
> Fix For: 1.6
>
>
> One of important disadvantages for yardstick framework that it requires to be copied on all hosts where server nodes will run and then server/gc logs can be found on these hosts only .
> But the script bin/benchmark-run-all.sh takes the configuration file where server hosts already defined (config/*.properties). Thus is it possible to add a first step for deploying yardstick on server hosts and then grab logs and copy into local yardstick directory.
> It will significantly reduce time to run benchmarks and get the results faster.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)