Struct: ComputeResourcePreference
Key | Field | Type | Description | Requiredness | Default value |
1 | computeResourceId | string | | required | |
2 | overridebyAiravata | bool | | required | true |
3 | loginUserName | string | | optional | |
4 | preferredJobSubmissionProtocol | compute_resource_model.JobSubmissionProtocol | | optional | |
5 | preferredDataMovementProtocol | data_movement_models.DataMovementProtocol | | optional | |
6 | preferredBatchQueue | string | | optional | |
7 | scratchLocation | string | | optional | |
8 | allocationProjectNumber | string | | optional | |
9 | resourceSpecificCredentialStoreToken | string | | optional | |
Gateway specific preferences for a Computer Resource
computeResourceId:
Corelate the preference to a compute resource.
overridebyAiravata:
If turned true, Airavata will override the preferences of better alternatives exist.
loginUserName:
If turned true, Airavata will override the preferences of better alternatives exist.
preferredJobSubmissionProtocol:
For resources with multiple job submission protocols, the gateway can pick a preferred option.
preferredDataMovementProtocol:
For resources with multiple data movement protocols, the gateway can pick a preferred option.
preferredBatchQueue:
Gateways can choose a defualt batch queue based on average job dimention, reservations or other metrics.
scratchLocation:
Path to the local scratch space on a HPC cluster. Typically used to create working directory for job execution.
allocationProjectNumber:
Typically used on HPC machines to charge computing usage to a account number. For instance, on XSEDE once an
allocation is approved, an allocation number is assigned. Before passing this number with job submittions, the
account to be used has to be added to the allocation.
resourceSpecificCredentialStoreToken:
Resource specific credential store token. If this token is specified, then it is superceeded by the gateway's
default credential store.