Workload Manager Server Settings
Overview of the Workload Manager Server settings that can be viewed and altered.
The following is a list of Workload Manager server settings available:
- Server Log Level
- An integer representation of a bit string that specifies all the Workload Manager
Server events to be logged. Default is 511. For further information about Workload
Manager Log Event Classes see the
PBS Professional Administrator's Guide.Note: Control does not currently support hexadecimal values.
- Mail From Address
- The username from which server-generated mail is sent to users.
- License Server
- The location where the Workload Manager looks for the license file and/or license server(s).
- Job History Duration
- A period of time expressed as [[HH:]MM:]SS[.milliseconds] that specifies the length of time PBS will keep each job's history. Default is two weeks.
- Job History
- Enabling this setting enables job history management.
- License Max
- Maximum number of licenses to be checked out at any time, i.e the maximum number of CPU licenses to keep in the PBS local license pool. It sets a cap on the number of CPUs that can be licensed at one time.
- License Min
- Minimum number of CPUs to permanently keep licensed, i.e. the minimum number of CPU licenses to keep in the PBS local license pool. This is the minimum number of licenses to keep checked out. If set to zero or unset, it is automatically set to the value of 1.
- License Linger Time
- The number of seconds to keep an unused CPU license, when the number of licenses is above the value given by License Min.
- Reservations
- Specifies whether or not advance and standing reservations can be created in this complex. When enabled, new reservations can be created. When disabled, new reservations cannot be created, but existing reservations are honored.
- Job Requeue Timeout
- The length of time that can be taken while requeueing a job. The minimum allowed value is 1 second and the maximum allowed value is 3 hours. Default is 45 seconds.
- Requeue Jobs on Vnode Failure
- Controls whether running jobs are automatically requeued or are deleted when the primary execution vnode fails. Number of seconds to wait after losing contact with Mother Superior before requeueing or deleting jobs.