[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] stand alone checkpointing
- Date: Thu, 12 Aug 2004 17:09:43 -0500
- From: Miron Livny <miron@xxxxxxxxxxx>
- Subject: Re: [Condor-users] stand alone checkpointing
I will let others answer the checkpointing question. Did you consider
running this program as a Condor "scheduler universe" job? It sounds as if
your program is responsible for managing Condor jobs like the DAGMan we
provide. The scheduler universe was designed for this kind of jobs.
At 12:17 PM 8/12/2004 +0100, you wrote:
I'm using a c program to submit jobs to condor as well as updating and
extracting information from a mysql database. The program won't be running
on condor but I understand that it's possible to compile this program
using condor_compile to enable checkpointing so that SIGUSR2 signal etc.
will not completely cripple all that's been achieved so far and we can
restart the program from the same place using condor_resume. Is this going
to work? Also I have the c line sleep(time_period) in my code which I
understand is not allowed when run on condor in the non-vanilla universe
but would it be ok compiling but not running on condor?
Condor-users mailing list