Cluster queues determine the machine(s) on which jobs will run.
Every multi-node ARC setup has a default cluster queue that can
submit jobs to any machine(s) in the cluster, and a local cluster queue
for submitting jobs to the local machine. If you would like certain types of jobs to be targeted
to specific machines, you can create additional cluster queues to address this desired
behavior.
You can use the arcconfig queue command in the ARC command directory to add, remove or modify ARC cluster queues.
In this section:
To add a cluster queue to an Ansys RSM Cluster (ARC), run the following command in the
ARC command directory, replacing
<queueName> with the desired queue name:
arcconfig queue add <queueName>
Once you have added a queue you can assign machines to it, enable it, and specify other custom settings. For options see Modifying a Cluster Queue.
Note: Newly added queues are not enabled upon creation.
To remove a cluster queue from an Ansys RSM Cluster (ARC), run the following command in
the ARC command directory, replacing
<queueName> with the name of the queue to be removed.
arcconfig queue remove <queueName>
To modify the settings of an existing cluster queue, such as the list of machines on which it will run jobs, use the following command in the ARC command directory, appending the desired options and/arguments to the command:
arcconfig queue modify <queueName>
For example, to enable a queue that you have created, run the following command:
arcconfig queue modify <queueName> -e true.
Note: Modifying the built-in default and
local cluster queues is not recommended. Even if you were to modify
the default queue, it would continue to automatically add new
machines.
Table 4.4: Options for 'arcconfig queue modify'
| Option | Usage | Format/Sample Value |
|---|---|---|
| -m <machines> | Enter the names of the machines to which the queue will submit jobs. | machine1:machine2:machine3 |
| -p <priority> | Specify the priority that this queue's jobs will be given in relation to other queues. Jobs will be run from the highest priority queue first. It is common to create a higher priority queue for smaller jobs so that they are processed before running large jobs that tie up computing resources for a long period of time. | -255 to 255, where -255 is highest priority, and 255 is lowest priority |
| -e <enabled> | Specify whether or not the queue is active (available for job submission). | True or False |
| -s <suspended> | You may want to suspend a queue temporarily in order to perform maintenance. When a queue is suspended, jobs can be submitted to the queue, but will not be processed until the suspension is lifted. | True or False |
| -n <numJobs> | Specify the maximum number of jobs that can be run from this queue. | 0-255 or * for no limit |
| -b <startTime> | Use this in conjunction with <endTime> to specify the time range in which jobs can be submitted using this queue. Defining an availability range can be useful when execution nodes or application licenses are only available at certain times of the day. | 00:00:01 |
| -q <endTime> | Use this in conjunction with <startTime> to specify the time range in which jobs can be submitted using this queue (see above). | 23:59:59 |
| -u <userList> | Specify which users you are allowing to use the queue. | all | user1:user2:userN |