[jira] [Commented] (AIRAVATA-3577) Gateway users to be able to add nodes and CPUs disregarding the CPUs per node auto calculation

["Dinuka Subashana De Silva (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/AIRAVATA-3577?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17500053#comment-17500053 ] 

Dinuka Subashana De Silva commented on AIRAVATA-3577:
-----------------------------------------------------

The changes are done and the pull request [1] is to be reviewed and merged. The change is applied to web components as well so, it will be reflected to whoever the application uses the queue setting component. Eg. amp gateway.

Additionally, when it comes to lock state from unlock state, it revalidates based on the node count. Please let me know if that has to be changed.

[1] [https://github.com/apache/airavata-django-portal/pull/69]

 

> Gateway users to be able to add nodes and CPUs disregarding the CPUs per node auto calculation
> ----------------------------------------------------------------------------------------------
>
>                 Key: AIRAVATA-3577
>                 URL: https://issues.apache.org/jira/browse/AIRAVATA-3577
>             Project: Airavata
>          Issue Type: New Feature
>          Components: Django Portal
>            Reporter: Eroma
>            Assignee: Dinuka Subashana De Silva
>            Priority: Major
>
> Currently in SciGaP we configure the central compute resource properties. In this configuration we also configure per partition, CPUs per node. When this configuration in place, when users are creating experiments, when user enters the number of nodes the CPUs will be auto calculated, when user changes the CPUs the nodes will be auto configured. 
> Lets say for shared partition the CPUs per node is configured as 10.
> In create experiment, if user adds 2 nodes the CPUs will change to 20. If user changes the CPUs per 8 the nodes will change to 1.
>  
> With this new development, we are giving the choice of disregarding this feature, to the user. If user needs, can remove this auto calculation and add the number of nodes and CPUs as needed for the computation.
> When experiments are shared if the shared user only has read access, then should not be able to change the original details when it comes to nodes and CPUs. If with write or higher permission user should be ablate change the nodes and CPUs disregarding the CPUs per node configuration.
> This should be available for new cloned experiments and in edit experiment as well.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)