You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by "Bernardo Gomez Palacio (JIRA)" <ji...@apache.org> on 2013/11/15 23:57:20 UTC
[jira] [Updated] (MESOS-813) As a CLI User I should be able get the
available resources of a Mesos Cluster.
[ https://issues.apache.org/jira/browse/MESOS-813?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Bernardo Gomez Palacio updated MESOS-813:
-----------------------------------------
Description:
Rationale:
As a CLI User I should be able to know how much memory/resources the Cluster has available before activating a framework. The Use Case arises when several CLI Users are interacting with Shark while the Cluster is already running additional frameworks. Ideally a CLI Users knows the available memory such that they can set a sensitive SPARK_MEM value. This will help reduce the number of Shark Queries hanging waiting for resources that they not necessarily require.
Implementation suggestion.
Will be simple to follow the mesos-ps script example to read the master/stats.json endpoint to present such metrics. The script could be mesos-stats and receive the specific metric you are interested in, should default to all.
Potential enhancement could be an argument that loops the command such that we have a continuos feed of the stats in the cli.
was:
Rationale:
As a CLI User I should be able to know how much memory/resources the Cluster has available before activating a framework. The Use Case arises when several CLI Users are interacting with Shark while the Cluster is already running additional frameworks. Ideally the CLI Users knows the available memory such that they can set a sensitive SPARK_MEM value. This will help reduce the number of Shark Queries hanging waiting for resources that they not necessarily require.
Implementation suggestion.
Will be simple to follow the mesos-ps script example to read the master/stats.json endpoint to present such metrics. The script could be mesos-stats and receive the specific metric you are interested in, should default to all.
Potential enhancement could be an argument that loops the command such that we have a continuos feed of the stats in the cli.
> As a CLI User I should be able get the available resources of a Mesos Cluster.
> ------------------------------------------------------------------------------
>
> Key: MESOS-813
> URL: https://issues.apache.org/jira/browse/MESOS-813
> Project: Mesos
> Issue Type: Improvement
> Reporter: Bernardo Gomez Palacio
> Priority: Minor
> Labels: cli
>
> Rationale:
> As a CLI User I should be able to know how much memory/resources the Cluster has available before activating a framework. The Use Case arises when several CLI Users are interacting with Shark while the Cluster is already running additional frameworks. Ideally a CLI Users knows the available memory such that they can set a sensitive SPARK_MEM value. This will help reduce the number of Shark Queries hanging waiting for resources that they not necessarily require.
> Implementation suggestion.
> Will be simple to follow the mesos-ps script example to read the master/stats.json endpoint to present such metrics. The script could be mesos-stats and receive the specific metric you are interested in, should default to all.
> Potential enhancement could be an argument that loops the command such that we have a continuos feed of the stats in the cli.
--
This message was sent by Atlassian JIRA
(v6.1#6144)