[torqueusers] Problem with excessive and incorrect "Begun execution" mails

Åke Sandgren ake.sandgren at hpc2n.umu.se
Wed Dec 21 10:40:22 MST 2005


On Wed, 2005-12-21 at 09:06 -0800, Garrick Staples wrote:
> We should do this on the torquedev list.
> 
> On Wed, Dec 21, 2005 at 02:48:49PM +0100, ?ke Sandgren alleged:
> > @@ -558,12 +558,13 @@ static void stat_update(
> >          if (pjob->ji_momstat == 0)  
> >            {
> >            /* first save since running job, */
> >            /* must save session id	   */
> >  
> >            job_save(pjob,SAVEJOB_FULL);
> > +	  svr_mailowner(pjob,MAIL_BEGIN,MAIL_NORMAL,NULL);
> >            }
> >  
> >          pjob->ji_momstat = time_now;
> >          }
> >  
> >        pstatus = (struct brp_status *)GET_NEXT(pstatus->brp_stlink);
> 
> Does this still work with really long prologs?  If prolog takes 4.5
> minutes to run, are there still dup BEGIN emails?

Don't know haven't tested that yet. Will try later.



More information about the torqueusers mailing list