[torqueusers] strange netload-output from 'pbsnodes -a'

Thomas Dargel td at chemie.hu-berlin.de
Thu Dec 1 23:53:42 MST 2005


> > > > ....,loadave=8.50,netload=18446744073701380569,state=....
> > > >                           ^^^^^^^^^^
> > > >                it starts with the same 10 digits!
> > > > 
> > > > Please, can anybody give me an explanation for that?
> > > 
> > > The netload variables are overflowing.  This patch should fix you up.
> > > 
> > 
> > This patch has fixed the problem.
> > I have another question concerning the format of the 'pbsnodes -a'-output:
> > 
> > How much affort is needed to switch the output values for totmem, availmem
> > and physmem from 'kb'- to 'GB'-format? I think this looks more 'human-readable'.
> > 
> 
> Maybe a -h for "human" like GNU df and du?

That's exactly what I mean, if it's possible, also for 'rectime';
the output for netload I modified to MBs:

...b,ncpus=8,loadave=0.00,netload=1281.1MB,state=free,rectim...

with the following changes in src/resmom/linux/mom_mach.c
near line 3905 (1.2.0p6):

/*  sprintf(ret_string,"%lu", */
/*    bytesRX[MAX_INTERFACES] + bytesTX[MAX_INTERFACES]); */
  sprintf(ret_string,"%.1lfMB",
    ( bytesRX[MAX_INTERFACES] + bytesTX[MAX_INTERFACES] ) / ( (double) 1024576 ) );

Not elegant, but it seems that it's work..


> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers


More information about the torqueusers mailing list