Can you check to see if you have multiple versions of LAM installed? A
common problem that has arisen recently is when the Linux distro has one
version of LAM installed (perhaps by default), but then a secondary
installation is located elsewhere (e.g., /usr/local). In this case, for
example, you may have multiple mpirun's in your PATH (e.g.,
/usr/bin/mpirun and /usr/local/bin/mpirun).
It can be easy to get confusing results in this case.
If this is not the case, let's investigate further.
On Thu, 11 Dec 2003, Becker, Robert P wrote:
> Konrad,
> lamboot hostfile was initiated prior to running any command. I have tried lamhalt multiple times which succeeds in shutting down the lamd daemon. Like I said, I am at a total loss to why this is happening. LSTC says the message I am getting is from the MPI, yet I can't see why it would since it passes every other test I can throw at it.
> -Rob Becker
>
> -----Original Message-----
> From: Konrad Karczewski [mailto:xeno_at_[hidden]]
> Sent: Thursday, December 11, 2003 2:37 AM
> To: General LAM/MPI mailing list
> Subject: Re: LAM: It says lamd isn't running?
>
>
>
>
> On Wed, 10 Dec 2003, Becker, Robert P wrote:
>
> > We are using a software package called LS-DYNA to run on our cluster.
> ....
> > By doing a ps -aux I am able to see that lamd is currently running. I
> > am also able to use the tping command and get a reply from both the
> > remote and the local node. LSTC, the maker of LS-DYNA is telling me
> > that the issue is with LAM/MPI and not their software. Since the test
> > suite runs properly and everything else is working I tend to disagree
> > with them, but I figured I would ask here to see if anyone else has had
> > any problems.
>
> It's not sufficient to have any lam daemon running, you have to start
> "your own" process - I suppose that you're not running lamboot before
> starting your task. Try running lamhalt (it should fail in this case) and
> then lamboot <lamhosts> as the user which is trying to run the
> application. If this doesn't help - you found an interesting case ;)
>
> best regards
> Konrad Karczewski
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
