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

Re: [Condor-users] DAG of DAGs



Thank you Greg

Can I provide from the outer dag to the Job submission file (in the inner.dag) a var?
So I will have something like this:

outer.dag:

SUBDAG EXTERNAL A inner.dag
SCRIPT POST A shouldContinue.sh
RETRY A 1000000
VAR procid = $some_env_set_that_the_shouldContinue.sh_provide


inner.dag:

Job  A  A.condor
Job  B  B.condor
Job  C  C.condor
Job  D  D.condor
PARENT A CHILD B PARENT B CHILD C
PARENT C CHILD D

A.Condor:

 Executable     = foo                                                   
 Requirements   = Memory >= 32 && OpSys == "IRIX65" && Arch =="SGI"    
 Rank         = Memory >= 64
 Image_Size     = 28 Meg                                                
 Arguments = procid)
 Error   = err.$(procid)                                               
 Input   = in.$(procid)                                                
 Output  = out.$(procid)                                               
 Log = foo.log

 Queue 150


Thanks
Sassy






On Thu, Jul 29, 2010 at 3:53 PM, Greg Thain <gthain@xxxxxxxxxxx> wrote:
Sassy Natan wrote:
Hi Group
 I was wonder if any one manage to do the following Task:
 I have a 4 Job definitions, which I have implement to run in a DAG configuration.
It looks like that:
 Job  A  A.condor
Job  B  B.condor
Job  C  C.condor
Job  D  D.condor
PARENT A CHILD B PARENT B CHILD C
PARENT C CHILD D
 The Job runs OK, but what I would like to do is to run this DAG in a loop configuration.

Basically, the only way to loop in DAGman is to mark a single node as "failed" via a post-script, and set a retry count.  This only loops that one node.  However, you can accomplish want you want by using nested dags.  By putting the above dag in its own file, say, inner.dag, you can then write an outer dag that looks like:

SUBDAG EXTERNAL A inner.dag
SCRIPT POST A shouldContinue.sh
RETRY A 1000000

and as long as the shell script (or program) shouldContinue.sh returns non-zero (up to the retry count), dagman will re-run the inner.dag.

-Greg

_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/