[MPICH] No route to host problem

[Ralph M. Butler]

Hi:

I can not say for sure what the problem here is.  Typically, a
'no route to host' problem is a network problem outside the
control of mpd/mpich2.  For configuration problems, we generally
suggest that folks begin debugging with the mpdcheck program
as described in the troubleshooting section.

--ralph

> > Date: Thu, 30 Jun 2005 16:05:34 -0500
> > From: Rajeev Thakur <thakur at mcs.anl.gov>
> > To: 'Ralph M. Butler' <rbutler at mtsu.edu>
> > Cc: 'Rusty Lusk' <lusk at mcs.anl.gov>
> > Subject: FW: [MPICH] No route to host problem
> >
> > Ralph,
> >       Any idea what this is? You can reply directly to him
> with a cc to the
> > list.
> >
> > Thanks,
> > Rajeev
> >
> > -----Original Message-----
> > From: owner-mpich-discuss at mcs.anl.gov
> > [mailto:owner-mpich-discuss at mcs.anl.gov] On Behalf Of
> Patrick E. Kane
> > Sent: Thursday, June 30, 2005 9:52 AM
> > To: mpich-discuss at mcs.anl.gov
> > Subject: Re: [MPICH] No route to host problem
> >
> > Below is more info on the "No route to host" problem.
> > I did an "strace" of the python process as it gave this error msg:
> >
> >    0: aborting job:
> >    0: Fatal error in MPI_Bcast: Other MPI error, error stack:
> >    0: MPI_Bcast(827): MPI_Bcast(buf=0xbffff88c, count=1,
> MPI_INT, root=0,
> >
> > Here an interesting part of the Linux system call trace log:
> >  ...
> >  connect(7, {sa_family=AF_INET, sin_port=htons(32805),
> >          sin_addr=inet_addr("192.168.9.3")},
> >          16) = -1 EINPROGRESS (Operation now in progress)
> >  poll([{fd=4, events=POLLIN}, {fd=7, events=POLLOUT,
> >       revents=POLLERR|POLLHUP}], 2, -1) = 1
> >  getpeername(7, 0xbffff480,
> >              [16])  = -1 ENOTCONN (Transport endpoint is
> not connected)
> >  getsockopt(7, SOL_SOCKET, SO_ERROR, [113], [4]) = 0
> >  write(2, "aborting job:\nFatal error in MPI_Bcast: Other
> MPI error"...