cluster properties#

The cluster.* properties are ONLY relevant if pipeline.env=cluster.

BioLockJ was originally designed to optimize effeciency on a cluster system, specifically one with a torque scheduler.

We recomend chaining configuration properties across multiple files. The cluster.* properties would go in the configuration file for you environement.

Property	Description
cluster.batchCommand	string Terminal command used to submit jobs on the cluster default: null
cluster.host	string The remote cluster host URL (used for ssh, scp, rsync, etc) default: null
cluster.jobHeader	string Header written at top of worker scripts default: null
cluster.modules	list List of cluster modules to load at start of worker scripts default: null
cluster.prologue	string To run at the start of every script after loading cluster modules (if any) default: null
cluster.returnsBatchIds	boolean Options Y/N. Does the cluster.batchCommand return a job id. If Y, if a job is submitted and no id is returned, that will be treated as a failure and the pipeline will stop. default: Y
cluster.statusCommand	string Terminal command used to check the status of jobs on the cluster default: null

For example, the following values are used for a torque scheduler in the file: cluster.properties

pipeline.env=cluster                   
cluster.batchCommand=qsub -q copperhead                   
cluster.host=hpc.uncc.edu                   
cluster.statusCommand=qstat                   

Each project's indicidule configuration file includes
pipeline.defaultProps = cluster.properties