Skip to content

User Settable Reservations

User settable reservations allows users to reserve compute nodes in advance on DataStar (P655) and the TeraGrid IA-64 cluster. Currently there are two ways to create user settable reservations. A user can use the command line tool, user_set_res, or they can take advantage of the SDSC user portal. Below are brief descriptions of the the two tools and how they can be used.

user_set_res

'user_set_res' tool is a command line tool used to create reservations on the IA64 cluster and Datastar. This tool is used by both the grid portal and GUR tools when they make reservations. This tool will allow you to request the number of nodes, the account that will be charged, the length of the reservation and the time frame in which you would like to have the nodes, as some of it's features. Users will need to manually identify available compute resources using the command and fill in the usage options described below.


    >user_set_res --account=<account>
    --nodes=<number of nodes>
    --duration=<duration(sec)>
    --earliest_start=<earliest start time(HH:MM_mm/dd/yyyy or epoch seconds)>
    --latest_end=<latest end time(HH:MM_mm/dd/yyyy or epoch seconds)>
    --email=<email address>
    [--sharedmap=<1#type:node_shared#cpu:1+memory:1>]
    [--featurelist=<comma-delimited list of node features>]
    [--qoslist=<comma-delimited list of allowed QOS>]

user_cancel_res

'user_cancel_res' tool is a command line tool used to cancel reservations on the IA64 cluster and Datastar.


    >user_cancel_res <reservation id>>

SDSC User Portal

For even more abstraction, if you prefer a web interface the SDSC portal will guide you through the process. From the portal, you can see which nodes are available in an easy calendar format. You may choose the number of nodes, the start time, and the duration of the reservation in a Web form and get an immediate confirmation without e-mail to the help desk. Some of the advantages are predictable work schedules, allows for effecient debugging, and running multiple small jobs without waiting in the queue