Hi folks,

First of all, I really love the new job configuration style. It makes it very easy to change swaths of jobs and to streamline the way jobs are submitted to the cluster. I had no issues with the dynamic runner and I like the way that tags interact. My question is for the dynamic job runner; is it possible to send the job to a different job *handler* dynamically? Or is it already given to the default handler at the time that the dynamic function is called? I want to do something like this:

def ncbi_blastn_wrapper(job):
    # Send to a different handler?
    inp_data = dict( [ ( da.name, da.dataset ) for da in job.input_datasets ] )
    inp_data.update( [ ( da.name, da.dataset ) for da in job.input_library_datasets ] )
    query_file = inp_data[ "query" ].file_name
    query_size = os.path.getsize( query_file )
    if query_size > 1024 * 1024:
       handlername = "bighandler"
   else:
       handlername = "smallhandler"
    return JobDestination(runner="pbs", handler=handlername, params={"Resource_List": "24:00:00"})

The way I have it set up now, my handlers are pointing at different clusters with different capabilities, and it would be cool to make the decision which to run at runtime.

Thanks a ton!

Carrie Ganote