[Simgrid-user] Fwd: exception handling when host is down

ashish lucknow luck.ashish at gmail.com
Fri Oct 2 10:43:24 CEST 2009


Any one have MSG Example that handles this kind of exception because i am
sure how many this i have to  free (received task by destroying it and (in
which functin ckptserver or at sender side also)) ples.. give full example
if posible. (i am new in exception handeling )


On Fri, Oct 2, 2009 at 1:41 PM, Ghislain Charrier <
ghislain.charrier at ens-lyon.fr> wrote:

> You just have to free the structures you allocated.
>
> On Fri, Oct 2, 2009 at 10:08 AM, ashish lucknow <luck.ashish at gmail.com>
> wrote:
> > The process should cleanup and then return from its main function ASAP.
> It
> > shouldn't retry
> > running something after this exception.
> >
> > how to do this cleanup in my code pleas give some more hint. t
> > Thanks..
> >
> >
> > On Fri, Oct 2, 2009 at 1:21 PM, Martin Quinson <martin.quinson at loria.fr>
> > wrote:
> >> Like I said, you shouldn't catch and shallow this exception. The point
> >> is that the process is running on a dead host. The process should
> >> cleanup and then return from its main function ASAP. It shouldn't retry
> >> running something after this exception. There is only 2 reason why we
> >> don't kill the process from the library itself.
> >> - so that you can free all the memory you malloced in this process
> >> - because the autorestart mechanism, supposed to reboot your process
> >> when the host comes back up, is not working atm. This should be better
> >> in the svn, and in the upcomming 3.3.4 (that I hope to release today).
> >>
> >> Bye, Mt.
> >>
> >> Le vendredi 02 octobre 2009 à 12:52 +0530, ashish lucknow a écrit :
> >> >
> >> >
> >> > ---------- Forwarded message ----------
> >> > From: ashish lucknow <luck.ashish at gmail.com>
> >> > Date: Fri, Oct 2, 2009 at 11:01 AM
> >> > Subject: Re: [Simgrid-user] exception handling when host is down
> >> > To: Ghislain Charrier <ghislain.charrier at ens-lyon.fr>
> >> >
> >> >
> >> > still unable to execute my program I have a ckpclint() that sends
> >> > dummy task to ckpserver () and the Host3 that is attached to ckpserver
> >> > will be
> >> > DOWN at time 1
> >> >
> >> >
> >> > program is like
> >> > int ckpclint(int argc,char *argv[])
> >> > {
> >> > - - - --  -
> >> >  -- - -- -
> >> > while(1){
> >> >            for(i=0;i<number_of_chunck;i++)
> >> >            {
> >> >                 TRY{
> >> >
> >> > ret=MSG_task_put(MSG_task_create("test",0,0,NULL),server[i %
> >> > localservercount],PORT_22);
> >> >                   INFO1("ret %d",ret);
> >> >
> >> >              CATCH(e){
> >> >                       ret= MSG_task_cancel(); // last task cancel with
> >> > para..
> >> >               INFO0("exception in sending task");
> >> >                     xbt_ex_free(e);
> >> >              }
> >> >    }
> >> >   }
> >> >
> >> >
> >> > int ckpserver(int argc, char *argv[])
> >> > {
> >> >    INFO0("IN SERVER");
> >> >     ....................
> >> >    TRY{
> >> >     MSG_task_get(&(task), PORT_22);
> >> >     INFO0("after");
> >> >      }
> >> >     CATCH(e)
> >> >     {
> >> >         //xbt_ex_display(&e);
> >> >         printf("execpti on taskget\n");
> >> >        printf("%s\n", e.msg);
> >> >         RETHROW;
> >> >       }
> >> >
> >> >     xbt_assert0(res == MSG_OK, "MSG_task_get failed");
> >> >     INFO1("Received \"%s\"", MSG_task_get_name(task));
> >> >     if (!strcmp(MSG_task_get_name(task), "FINALIZE1")) {
> >> >       MSG_task_destroy(task);
> >> >       break;
> >> >     }
> >> >
> >> >     INFO1("Processing \"%s\"", MSG_task_get_name(task));
> >> >     TRY{
> >> >     MSG_task_execute(task);
> >> >    }
> >> >   CATCH(ee)
> >> >   {
> >> >  MSG_task_destroy(task);
> >> >        RETHROW;
> >> >    }
> >> > -..........
> >> > }
> >> >
> >> > OUT ......
> >> >
> >> >
> >> > [Host0:ckpclint:(1) 1.000000] [msg_test/INFO] Sent
> >> > [Host0:ckpclint:(1) 1.000000] [msg_test/INFO] Sending "chunck_2" to
> >> > "Host3"
> >> > [Host0:ckpclint:(1) 1.000000] [msg_test/INFO] count 51
> >> > [Host3:ckpserver:(4) 1.000000] [msg_test/INFO] after
> >> > [Host3:ckpserver:(4) 1.000000] [msg_test/INFO] Received "chunck_1"
> >> > [Host3:ckpserver:(4) 1.000000] [msg_test/INFO] Processing "chunck_1"
> >> > task execute execption
> >> >
> >> > ** SimGrid: UNCAUGHT EXCEPTION received on Host3(4): category:
> >> > unknown_err; value: 0
> >> > ** Host failed, you cannot call this function. (state=0)
> >> > ** Thrown by ckpserver() in this process
> >> > [Host3:ckpserver:(4) 1.000000] xbt/ex.c:117: [xbt_ex/CRITICAL] Host
> >> > failed, you cannot call this function. (state=0)
> >> >
> >> > **   In MSG_task_execute()
> >> > at /home/luck/temp/simgrid-3.3.3/src/msg/gos.c:47
> >> > **   In ckpserver() at ??:0
> >> > **   In smx_ctx_sysv_stop()
> >> > at /home/luck/temp/simgrid-3.3.3/src/simix/smx_context_sysv.c:145
> >> > **   In makecontext() at ??:0
> >> >
> >> > **   In double_update()
> >> > at /home/luck/temp/simgrid-3.3.3/src/../src/include/surf/maxmin.h:18
> >> >
> >> > **   In surf_solve()
> >> > at /home/luck/temp/simgrid-3.3.3/src/surf/surf.c:414
> >> > **   In SIMIX_solve()
> >> > at /home/luck/temp/simgrid-3.3.3/src/simix/smx_global.c:305
> >> > **   In MSG_main()
> >> > at /home/luck/temp/simgrid-3.3.3/src/msg/global.c:161
> >> > **   In test_all() at ??:0
> >> > Aborted
> >> >
> >> >
> >> >
> >> > How to stop this Aborting...
> >> >
> >> >
> >> > Ashsih Kumar
> >> >
> >> >
> >> >
> >> >
> >> > --
> >> > Ashsih Kumar
> >> > _______________________________________________
> >> > Simgrid-user mailing list
> >> > Simgrid-user at lists.gforge.inria.fr
> >> > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/simgrid-user
> >> -- > Strengths: What are the major reasons to accept the paper? I did
> >> not find strengths; I stopped reading the paper on page 12, so I may
> >> have missed something. -- Bastard Reviewer From Hell
> >>
> >
> >
> >
> > --
> > Ashsih Kumar
> >
> > _______________________________________________
> > Simgrid-user mailing list
> > Simgrid-user at lists.gforge.inria.fr
> > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/simgrid-user
> >
>
>
>
> --
> Ghislain Charrier
> Doctorant INRIA                          Équipe GRAAL
> Bureau 174 UCBL Gerland           Tel: +33 4 37 28 76 45
> http://graal.ens-lyon.fr/~gcharrie <http://graal.ens-lyon.fr/%7Egcharrie>
> Laboratoire de l'Informatique du Parallélisme (LIP)
> École Normale Supérieure de Lyon
> 46 Allée d'Italie 69364 Lyon Cedex 07 - France
>



-- 
Ashsih Kumar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.gforge.inria.fr/pipermail/simgrid-user/attachments/20091002/be6ebe3e/attachment-0001.htm 


More information about the Simgrid-user mailing list