4.1.1 Configuring the request element for the Job Submit form
Viewpoint allows you to customize and configure the msub command to match the needs and requirements specific to your site. See the Moab Workload Manager documentation.
To configure the Job Submit request element
- Open the jobs.xml file located in the Viewpoint home directory. Locate the <request> element.
- Add any of the following child elements:
- <account> – This field corresponds to the Account attribute and is the name of account associated with job (It accepts any string value decider. See Adding a string decider.).
- <additional-attributes> – Any additional command line arguments to be appended to the end of the 'msub' request. Each argument should be defined using a <attribute> element. Each child <attribute> element has two parts:
- <value> (required) – The value of the command line argument (It accepts any string value decider. See Adding a string decider.).
- <condition> (optional) – If this is specified, Viewpoint includes the argument only if the condition evaluates to true (It accepts any Boolean decider. See Setting a Boolean decider.).
- <apply-hold> – This field corresponds to the Hold attribute and specifies that a user hold be applied to the job at submission time (It accepts any Boolean value decider. See Setting a Boolean decider.).
- <class> – This field corresponds to the Destination Queue (Class) attribute and defines the destination of the job. Moab uses the term Destination Queue for class (It accepts any string value decider. See Adding a string decider.).
- <error-path> – This field corresponds to the Error Path attribute and defines the path to be used for the standard error stream of the batch job (It accepts any string value decider. See Adding a string decider.).
- <execution-dir> – This field corresponds to the Execution Directory attribute and is the directory in which the job should execute (It accepts any string value decider. See Adding a string decider.).
- <generic-resources> – This field corresponds to the generic resources attribute. This works only with the table-map component. Here is an example:
<generic-resource>
<component id="generic_resource_table_map" />
</generic-resource>
- <gpu-count> – This field corresponds to the Resource List attribute and is the number of GPUs per node requested (It accepts any integer value decider. See Adding an integer decider.).
- <job-name> – This field corresponds to the Name attribute and is the user-specified job name (It accepts any string value decider. See Adding a string decider.).
- <job-script> – This field corresponds to the Job Script attribute(It accepts any string value decider. See Adding a string decider.). Typically, this value is determined from an upload-script component or a write-script component, however, you can hard code the value or determine the value from other means. Here is an example where the job script is determined from either an upload-script component or a write-script component, depending on which option the user had selected:
<job-script>
<calculate operation="conditional">
<option>
<value>
<component id="upload-script" />
</value>
<condition comparison="equal">
<first>
<component id="script-chooser" />
</first>
<second>Upload</second>
</condition>
</option>
<option>
<value>
<component id="write-script" />
</value>
<condition comparison="equal">
<first>
<component id="script-chooser" />
</first>
<second>Create New</second>
</condition>
</option>
</calculate>
</job-script>
- <memory> – This field corresponds to the Resource List attribute and is the maximum amount of physical memory used by the job (It accepts any integer value decider. See Adding an integer decider.).
- <node-count> – This field corresponds to the Resource List attribute and is the number of nodes to be reserved for exclusive use by the job (It accepts any integer value decider. See Adding an integer decider.).
- <node-features> – The list of node features the job requires (It accepts any string list value decider. See Adding a string decider.).
- <operating-system> – The administrator defined operating system (It accepts any string value decider. See Adding a string decider.).
- <output-path> – This field corresponds to the Output Path attribute and defines the path to be used for the standard output stream of the batch job (It accepts any string value decider. See Adding a string decider.).
- <partition> – The required partition (It accepts any string value decider. See Adding a string decider.).
- <preemptible> – The job is a preemptee and therefore can be preempted by other jobs (It accepts any Boolean value decider. See Setting a Boolean decider.).
- <priority> – This is the priority of the job (It accepts any integer value decider. See Adding an integer decider.). Normally, this is between -1024 and 0.
- <proc-count> – This field corresponds to the Resource List attribute and is the number of processors per node requested (It accepts any integer value decider. See Adding an integer decider.).
- <qos> – The Quality of Service associated with job (It accepts any string value decider. See Adding a string decider.).
- <restartable> – This field corresponds to the Rerunable attribute and declares whether the job is rerunable (It accepts any Boolean value decider. See Setting a Boolean decider.).
- <swap> – The maximum amount of virtual memory used by all concurrent processes in the job in MB which is the amount of swap disk required by job (It accepts any integer value decider. See Adding an integer decider.).
- <variables> – A collection of variables to be set on the job. This can have one or more <variable> child elements. Each child <variable> element has a minimum of two parts:
- <key> (required) – The name of the variable (It accepts any string decider. See Adding a string decider.).
- <value> (required) – The value of the variable (It accepts any string decider. See Adding a string decider.).
- <rule> (optional) – If this is specified, Viewpoint includes the variable only if the rule evaluates to true (It accepts any Boolean value decider. See Setting a Boolean decider.).
- <vm-usage> – The environment in which the job will be run. This value must be one of the following:
- requirepm – The job runs on a physical machine.
- prefpm – The job runs on a physical machine, if possible; if not, it runs on a virtual machine.
- createvm – The job creates a virtual machine on which to run. The newly-created VM is destroyed upon job completion.
- createpersistentvm – The job creates a virtual machine on which to run, but the newly-created VM persists after job completion.
- requirevm – The job does not run until a virtual machine is available on which it can run.
- prefvm – The job runs on a virtual machine, if possible; if not, it runs on a physical machine.
- <wall-time> – This field corresponds to the Resource List attribute and is the maximum amount of real time during which the job can be in the running state in seconds (It accepts any string integer decider. See Adding an integer decider.).
Related topics