[ 
https://issues.apache.org/jira/browse/VCL-170?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Aaron Peeler updated VCL-170:
-----------------------------
    Description: 
Changed to provide similar functionality by using the predictive reload 
modules. Create a Level_2.pm which returns unload / power off node. 

Major change as part of this is to move the predictive module setting from the 
management node to the computer table. This will allow for finer control over 
single or groups of computers. 

changes:
- add unload routine to each provisioning module
DB - move predictivemoduleid to computer table from managementnode table, 
update module table
Backend - create Level_2.pm.
- add an action return value to the array that is returned from the predictive 
modules. (unload, reload) 
- update code in Datastructure to pull predictive module info from request data 
hash, instead of management hash.
- frontend - add predictive reload module to computer information, remove from 
management node interface
- add code to the upgrade script to move the predictivemoduleid from the 
management node table to the computer table


Original:
Add support for general power management.  power off idle blades based on 
current usage levels

Through the web-interface allow site admins to define a policy based on some 
usage level value when and how many blades to shutdown. 

requires a new computer state "offavailable" or something like it

the backend would need to figure out the current load on the given system, 
either locally or through xmlrpc calls.

These are initial thoughts and should be discussed more

  was:
Add support for general power management.  power off idle blades based on 
current usage levels

Through the web-interface allow site admins to define a policy based on some 
usage level value when and how many blades to shutdown. 

requires a new computer state "offavailable" or something like it

the backend would need to figure out the current load on the given system, 
either locally or through xmlrpc calls.

These are initial thoughts and should be discussed more

        Summary: option to power off blades are reservation - reload module  
(was: Power Management)

> option to power off blades are reservation - reload module
> ----------------------------------------------------------
>
>                 Key: VCL-170
>                 URL: https://issues.apache.org/jira/browse/VCL-170
>             Project: VCL
>          Issue Type: New Feature
>          Components: database, vcld (backend), web gui (frontend)
>    Affects Versions: 2.2
>            Reporter: Aaron Peeler
>             Fix For: 2.4
>
>
> Changed to provide similar functionality by using the predictive reload 
> modules. Create a Level_2.pm which returns unload / power off node. 
> Major change as part of this is to move the predictive module setting from 
> the management node to the computer table. This will allow for finer control 
> over single or groups of computers. 
> changes:
> - add unload routine to each provisioning module
> DB - move predictivemoduleid to computer table from managementnode table, 
> update module table
> Backend - create Level_2.pm.
> - add an action return value to the array that is returned from the 
> predictive modules. (unload, reload) 
> - update code in Datastructure to pull predictive module info from request 
> data hash, instead of management hash.
> - frontend - add predictive reload module to computer information, remove 
> from management node interface
> - add code to the upgrade script to move the predictivemoduleid from the 
> management node table to the computer table
> Original:
> Add support for general power management.  power off idle blades based on 
> current usage levels
> Through the web-interface allow site admins to define a policy based on some 
> usage level value when and how many blades to shutdown. 
> requires a new computer state "offavailable" or something like it
> the backend would need to figure out the current load on the given system, 
> either locally or through xmlrpc calls.
> These are initial thoughts and should be discussed more



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to