[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] proposed change to DAG NOOP nodes

On 2/18/2016 5:04 PM, Dimitri Maziuk wrote:
> On 02/18/2016 02:45 PM, Kent WENGER wrote:
>> Well, the current implementation of NOOP nodes totally ignores the
>> submit file.  I guess we could add a knob that says either "use the
>> submit file" or "ignore the submit file".
> The reason I asked is because a while back I "temporarily" NOOPed a --
> perfectly valid at the time -- dag node. So I was wondering what you
> meant by "remove the don't actually submit code": if that node gets
> actually submitted at this point, who knows what may happen.

In Kent's original proposal at the start of this thread, he said that DAGMan will submit a NOOP node with condor_submit flag
     -append "noop_job = true"

Seems like this should take care of knowing what would happen in the case of a NOOP dag node having a valid submit file - what would happen is exactly as described in the man page for condor_submit for noop_job=true:

  noop_job = <ClassAd Boolean Expression>

  When this boolean expression is  True , the job is  immediately
  removed  from  the queue, and HTCondor makes no attempt at run-
  ning the job. The log file for the job will show a job  submit-
  ted  event  and a job terminated event, along with an exit code
  of 0, unless the user specifies  a  different  signal  or  exit

Am I missing something?  I don't think we need yet another knob here
to say ignore the submit file; in fact, I think such a knob would defeat part of the intent for changing how the current NOOP job mechanism works.