Recent versons of snow signal an error if the value returned from a
worker indicates an error. The error handling facilities in snow are
still evolving; for now id you don't want an error on a worker to
become an error an the master you need to catch the error in the
worker yourself and produce an appropriate result, e.g. by replacing
MCexe by something like
function(...) tryCatch(MCexe(...), error = function(e) NULL)
if NULL is OK as a result when the MCexe computation produces an error.
luke
On Fri, 27 Mar 2009, jgarcia at ija.csic.es wrote:
> Hello,
>
> I have a program that used to run well in October, it uses library snow.
> Since then, one change has ocurred (snow library has been updated) and
> another could have ocurred (I've unadvertently modified something).
>
> Anyway, now when I make the call:
>
> parallel.model.results <- clusterApply(cl,processors.struct,MCexe)
>
> exactly as I used to do, where MCexe is my function and processors.struct
> is a list containing everything required by MCexe, I obtain the following
> error:
>
> Error in checkForRemoteErrors(val) :
> 2 nodes produced errors; first error: incorrect number of dimensions
>
> Please, do you have any clue about what could be the error?
>
> Best regards,
>
> Javier Garc?a-Pintado
>
> _______________________________________________
> R-sig-hpc mailing list
> R-sig-hpc at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-hpc
>
--
Luke Tierney
Chair, Statistics and Actuarial Science
Ralph E. Wareham Professor of Mathematical Sciences
University of Iowa Phone: 319-335-3386
Department of Statistics and Fax: 319-335-3017
Actuarial Science
241 Schaeffer Hall email: luke at stat.uiowa.edu
Iowa City, IA 52242 WWW: http://www.stat.uiowa.edu