[torqueusers] What is cartwire.ncsa.uiuc.edu?

Lorenzo Campo lorenzo118 at interfree.it
Tue Oct 18 02:48:01 MDT 2005


Dave,
no, I'm using the source of torque 1.2.0p6, downloaded and compiled on my 
machines following the quickstart guide (and compiling with --with-scp 
option). I'm using the pbs_sched scheduler, no RPM or other binary at all. 
There is some issue in running jobs with more than one compute node?
Lorenzo




At 19.16 17/10/2005, you wrote:
>Lorenzo,
>
>   This string does not exist in the original distribution.  Are you
>using an RPM, debian, or other style binary build?  If you are using a
>binary distribution, do you know who packaged it?
>
>Dave
>
>On Mon, 2005-10-17 at 17:53 +0200, Lorenzo Campo wrote:
> > Hi all,
> > I installed and configure torque1.2.0p6 on my 16-processor Fedora Core 3
> > cluster, but a strange thong happens. All nodes are free (by pbsnodes -a)
> > but when I try to submit a job on 4 processors, the job try to run, the
> > reults in the .out file (the .err file is empty) are:
> >
> > p0_7109: (60.382862) Procgroup:
> > p0_7109: (60.382985)     entry 0: medusa003.dicea.unifi.it 0 0
> > /home/berna/./helloMpi berna
> > p0_7109: (60.383012)     entry 1: cartwire.ncsa.uiuc.edu 1 1
> > /home/berna/./helloMpi berna
> > p0_7109: (60.383035)     entry 2: cartwire.ncsa.uiuc.edu 1 2
> > /home/berna/./helloMpi berna
> > p0_7109: (60.383063)     entry 3: cartwire.ncsa.uiuc.edu 1 3
> > /home/berna/./helloMpi berna
> > p0_7109:  p4_error: Could not gethostbyname for host
> > cartwire.ncsa.uiuc.edu; may be invalid name
> > : 61
> >
> > Why it tries to connect to cartwire.ncsa.uiuc.edu (and WHAT IS
> > cartwire.ncsa.uiuc.edu, I never put such an address in none of my files in
> > my cluster)? With 2 processor (same program) I obtain:
> >
> > p0_9131: (60.758792) Procgroup:
> > p0_9131: (60.758893)     entry 0: medusa001.dicea.unifi.it 0 0
> > /home/berna/./helloMpi berna
> > p0_9131: (60.758916)     entry 1: cartwire.ncsa.uiuc.edu 1 1
> > /home/berna/./helloMpi berna
> > p0_9131:  p4_error: Could not gethostbyname for host
> > cartwire.ncsa.uiuc.edu; may be invalid name
> > : 61
> >
> >
> > and so on. What's going on? Where in the PBS is put this
> > "cartwire.ncsa.uiuc.edu" and how I can to avoid this? It seems that it
> > takes only the n-th processor of my cluster (the second if I request 2
> > processors, the third if I request 3, and so on) and then it tries to
> > submit to this external address.
> > Any idea?
> > Thank you
> > Lorenzo Campo
> >
> >
> > _______________________________________________
> > torqueusers mailing list
> > torqueusers at supercluster.org
> > http://www.supercluster.org/mailman/listinfo/torqueusers




More information about the torqueusers mailing list