Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: Re: [OMPI users] Cluster : received unexpected process identifier
From: Jeffrey Squyres (jsquyres_at_[hidden])
Date: 2012-04-04 10:47:25


Do you have multiple IP interfaces on a single machine on the same IP subnet, perchance?

If so, don't do that. :-) Put each of your IP addresses in a different IP subnet.

On Apr 3, 2012, at 10:30 PM, Rohan Deshpande wrote:

> Hi,
>
> I am running a MPI program using cluster and tcp communication.
>
> To run I am using: mpirun --prefix /usr/local/ --mca btl tcp,self --hostfile slaves -np 6 scatter
>
>
> I am getting following output:
>
> Process 0 on host host1 has elements 0 1 2 3
> Process 1 on host host1 has elements 4 5 6 7
> Process 2 on host host1 has elements 8 9 10 11
> Process 3 on host host1 has elements 12 13 14 15
> [slave1][[24708,1],5][btl_tcp_endpoint.c:486:mca_btl_tcp_endpoint_recv_connect_ack] received unexpected process identifier [[24708,1],4]
>
> When trying to communicate with slave1 i get received unexpected process identifier error.
>
> My PATH and LD_LIBRARY_PATH are correctly set in .bashrc file.
>
>
>
> Source code -
>
>
> #include <stdlib.h>
> #include <stdio.h>
> #include "mpi.h"
> #define MAXPROC 8 /* Max number of procsses */
> #define LENGTH 24 /* length of buffer */
>
> int main(int argc, char* argv[]) {
> int i, j, np, me;
> const int nametag = 40; /* Tag value for sending name */
> const int datatag = 45; /* Tag value for sending data */
> const int root = 0; /* Root process in scatter */
> MPI_Status status; /* Status object for receive */
>
> char myname[MPI_MAX_PROCESSOR_NAME]; /* Local host name string */
> char hostname[MAXPROC][MPI_MAX_PROCESSOR_NAME]; /* Received host names */
> int namelen;
>
> int x[LENGTH]; /* Send buffer */
> int y[LENGTH]; /* Receive buffer */
>
> MPI_Init(&argc, &argv); /* Initialize MPI */
> MPI_Comm_size(MPI_COMM_WORLD, &np); /* Get nr of processes */
> MPI_Comm_rank(MPI_COMM_WORLD, &me); /* Get own identifier */
>
> MPI_Get_processor_name(myname, &namelen); /* Get host name */
> myname[namelen++] = (char)0; /* Terminating null byte */
>
> /* Check that we have an even number of processes and at most MAXPROC */
> if (np>MAXPROC || np%2 != 0) {
> if (me == 0) {
> printf("You have to use an even number of processes (at most %d)\n", MAXPROC);
> }
> MPI_Finalize();
> exit(0);
> }
>
> if (me == 0) { /* Process 0 does this */
>
> /* Initialize the array x with values 0 .. LENGTH-1 */
> for (i=0; i<LENGTH; i++) {
> x[i] = i;
> }
>
> printf("Process %d on host %s is distributing array x to all %d processes\n\n",
> me, myname, np);
>
> /* Scatter the array x to all proceses, place it in y */
> MPI_Scatter(x, LENGTH/np, MPI_INT, y, LENGTH/np, MPI_INT, root,
> MPI_COMM_WORLD);
>
> /* Print out own portion of the scattered array */
> printf("Process %d on host %s has elements", me, myname);
> for (i=0; i<LENGTH/np; i++) {
> printf(" %d", y[i]);
> }
> printf("\n");
>
> /* Receive messages with hostname and the scattered data */
> /* from all other processes */
> for (i=1; i<np; i++) {
> MPI_Recv (hostname[i], namelen, MPI_CHAR, i, nametag, MPI_COMM_WORLD,
> &status);
> MPI_Recv (y, LENGTH/np, MPI_INT, i, datatag, MPI_COMM_WORLD, &status);
> printf("Process %d on host %s has elements", i, hostname[i]);
> for (j=0; j<LENGTH/np; j++) {
> printf(" %d", y[j]);
> }
> printf("\n");
> }
>
> printf("Ready\n");
>
>
> } else { /* all other processes do this */
>
> /* Receive the scattered array from process 0, place it in array y */
> MPI_Scatter(x, LENGTH/np, MPI_INT, y, LENGTH/np, MPI_INT, root, \
> MPI_COMM_WORLD);
> /* Send own name back to process 0 */
> MPI_Send (myname, namelen, MPI_CHAR, 0, nametag, MPI_COMM_WORLD);
> /* Send the received array back to process 0 */
> MPI_Send (y, LENGTH/np, MPI_INT, 0, datatag, MPI_COMM_WORLD);
>
> }
>
> MPI_Finalize();
> exit(0);
> }
>
> Any idea what could be wrong?
>
>
> Thanks
>
>
>
>
>
> _______________________________________________
> users mailing list
> users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/users

-- 
Jeff Squyres
jsquyres_at_[hidden]
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/