[torqueusers] Node np parameter adjusted automatically since 2.1.x ?

Garrick Staples garrick at clusterresources.com
Tue Jun 20 21:26:36 MDT 2006


On Tue, Jun 20, 2006 at 11:10:58PM -0400, Daniel Widyono alleged:
> > Just comment out the 'if else' block at lines 1085-1101 in
> > src/server/node_manager.c.
> 
> > The current logic is to only change np if MOM's ncpus is greater than
> > np.  Maybe a better logic is to only change if (np==1 && ncpus>np)?
> 
> I don't understand.  If the nodes file exists, does it not follow that the
> admin already set up the server, assumedly in a proper fashion?  Why alter
> something which the admin must normally set up manually?  It feels like an
> RPM install which clobbers a manually edited config file...

We're trying to get to an easier initial configuration.  For the first time
setup, I don't know if we can "assume" a proper setup.  That particular
config seems redundant to me anyways.  If MOM already reads the number of
CPUs, and advertises it to pbs_server, why shouldn't the config be automatic?

A server attribute, "auto_node_np", just landed upstream to control this behaviour.

I'm thinking along the lines of a tri-state value:
  Unset means "set np=ncpus if (np==1 && ncpus>np)"
  True is more strict with "set np=ncpus if (np!=ncpus)"
  False completely disables the feature.
  The default would be unset.

Does that sound reasonable?



More information about the torqueusers mailing list