[jira] [Commented] (OODT-231) BatchMgr, JobMonitor and JobQueue to interface with Torque

["Chris A. Mattmann (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/OODT-231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13051640#comment-13051640 ] 

Chris A. Mattmann commented on OODT-231:
----------------------------------------

originally reported by Cecilia S Cheng from JPL.

> BatchMgr, JobMonitor and JobQueue to interface with Torque
> ----------------------------------------------------------
>
>                 Key: OODT-231
>                 URL: https://issues.apache.org/jira/browse/OODT-231
>             Project: OODT
>          Issue Type: New Feature
>          Components: resource manager
>         Environment: from JPL's internal JIRA
>            Reporter: Chris A. Mattmann
>            Assignee: Chris A. Mattmann
>             Fix For: 0.4
>
>
> ACOS and OCO2 are attempting to run operations and their SCF on the same cluster. Currently, operations is using the OODT Resource Manager to run specific tasks on nodes assigned to operations. The SCF users use Torque to perform the same function. Depending on workload, nodes need to be switched between operations and the SCF. It is difficult and time consuming to change node allocations in the Resource Manager. It would be very easy to do if all nodes were managed by Torque. Torque also provides more powerful and more flexible resource management capabilities than the Resource Manager. In Torque, the operator can change allocation of nodes w/o bringing the whole system now; and Torque also provides more sophisticated queues such as priorities. Since not all the project has other jobs that are run outside of PCS, it's desirable to have 1 Torque manage both the PCS & non-PCS jobs.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira