Re: [HTCondor-users] Access to renamed DAG node names

["R. Kent Wenger" <wenger@xxxxxxxxxxx>]

On Mon, 12 Jan 2015, Brian Candler wrote:

> Actually a couple of other requirements:
>
> - a generic SCRIPT POST which reads the output of a job and delivers it 
> elsewhere
> - given a report of a failed node in JOBSTATE.LOG or NODE_STATUS file, be 
> able to locate the stdout and stderr relating to that job, for diagnostic 
> purposes

Hmm -- you're saying that for failed nodes you'd like to see the values of 
output and error put directly into the JOBSTATE.LOG or NODE_STATUS files,
so you could find them without having to look at the submit file?

> So I might had dag.dag containing
>
> SPLICE FOO dag1.dag
> SPLICE BAR dag2.dag
>
> with dag1.dag and dag2.dag both containing
>
> JOB A generic.sub
> VARS A job=A
> SCRIPT POST A send_results $(job)

For PRE and POST scripts, just use $job without the parentheses, so:

  SCRIPT POST A send_results $job

will work.

> and generic.sub has
>
> output = $(job).out
> error = $(job).err

For the actual HTCondor job to you have to use the DAG VARS feature as I 
mentioned in my previous email.

> The documentation for -usedagdir suggests that it's only when multiple 
> independent dags are submitted on the command line; it doesn't describe 
> whether or not this applies to subdags or splices.

It applies to sub-DAGs but not splices -- I guess the manual should be 
clearer.

Kent