The PATH variable contains /home/hx019035/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/maui/bin/:

/home/hx019035/bin contains the local installation of OMPI 1.3.3

 

The LD_LIBRARY_PATH variable contains /home/hx019035/lib:

 

These variables are being set in the .profile file on the hpcc00 node.

 

Would there be a change anywhere else?

 

 

From: Ralph Castain [mailto:rhc.openmpi@gmail.com] On Behalf Of Ralph Castain
Sent: 07 October 2009 13:32
To: Blesson Varghese
Subject: Re: [OMPI users] MPI_Comm_accept()/connect() errors

 

Yes, it does. But the error message indicates a 1.2 version is running on hpcc00.

 

On Oct 7, 2009, at 5:46 AM, Blesson Varghese wrote:

 

Just a quick question. Would mpirun –version give me the version of the mpirun being executed? I am getting the result of that as 1.3.3.

 

From: Ralph Castain [mailto:rhc.openmpi@gmail.com] On Behalf Of Ralph Castain
Sent: 07 October 2009 11:58
To: Blesson Varghese
Subject: Re: [OMPI users] MPI_Comm_accept()/connect() errors

 

Hate to tell you this, but your output clearly indicates you are NOT running 1.3.3 - that is an output from a 1.2.x version of OMPI.

 

Check you path and ld_library_path - you're still picking up the 1.2.5 version somewhere.

 

 

On Oct 7, 2009, at 4:05 AM, Blesson Varghese wrote:



Hi,

 

Please refer to the emails below.

 

I have made an upgrade to Open MPI 1.3.3 as suggested. The necessary environment variables have all been set. Attaching the output of ompi_info –all. However, the errors continue to persist.

 

[hpcc00:31864] [0,0,0] ORTE_ERROR_LOG: Not found in file dss/dss_unpack.c at line 209

[hpcc00:31864] [0,0,0] ORTE_ERROR_LOG: Not found in file communicator/comm_dyn.c at line 186

[hpcc00:31864] *** An error occurred in MPI_Comm_connect

[hpcc00:31864] *** on communicator MPI_COMM_WORLD

[hpcc00:31864] *** MPI_ERR_INTERN: internal error

[hpcc00:31864] *** MPI_ERRORS_ARE_FATAL (goodbye)

 

 

The server program is as follows:

 

#include <mpi.h>

#include <stdio.h>

#include <stdlib.h>

 

int main( int argc, char **argv )

{

      MPI_Comm client;

      MPI_Status status;

      char port_name[MPI_MAX_PORT_NAME];

      int buf;

      int size, again;

      MPI_Info portInfo;

 

      MPI_Init( &argc, &argv );

 

      MPI_Comm_size(MPI_COMM_WORLD, &size);

   

      MPI_Open_port(MPI_INFO_NULL, port_name);

 

      printf("server available at %s\n",port_name);

     

      MPI_Comm_accept(port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &client );

      MPI_Recv(&buf, 1, MPI_INT, MPI_ANY_SOURCE, MPI_ANY_TAG, client, &status );

      MPI_Comm_disconnect( &client );

}

 

The client program is as follows:

 

#include <mpi.h>

#include <stdlib.h>

#include <stdio.h>

#include <string.h>

 

int main( int argc, char **argv )

{

    MPI_Comm server;

    int buf = 8;

    char port_name[MPI_MAX_PORT_NAME];

    MPI_Info portInfo;

 

    MPI_Init( &argc, &argv );

 

    strcpy(port_name, "0.0.0:2000"); //The port name is hardcoded since 0.0.0:2000 is generated by the server program

    MPI_Comm_connect(port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &server );

 

    MPI_Send(&buf, 1, MPI_INT, 0, 1, server );

   

    MPI_Comm_disconnect( &server );

    MPI_Finalize();

    return 0;

}

 

Would you please advise?

 

Regards,

Blesson.

 

 

-----Original Message-----
From: Blesson Varghese [mailto:hx019035@reading.ac.uk] 
Sent: 03 October 2009 12:20
To: 'Jeff Squyres'
Subject: RE: [OMPI users] MPI_Comm_accept()/connect() errors

 

Thank you. I shall try the upgrade very soon.

 

-----Original Message-----

From: Jeff Squyres [mailto:jsquyres@cisco.com]

Sent: 03 October 2009 12:18

To: Blesson Varghese

Subject: Re: [OMPI users] MPI_Comm_accept()/connect() errors

 

On Oct 3, 2009, at 7:14 AM, Blesson Varghese wrote:

 

> Thanks for your reply Jeff. Since, it is a teaching cluster of the

> University,  I am quite unsure if I would be able to upgrade it very 

> soon.

> 

> Do you reckon that the error is due to the Open MPI version?

> 

 

You can always install your own version of Open MPI under your $HOME 

or somesuch -- there is no requirement that Open MPI is installed by 

root in a central location.

 

That being said, you might want to check with your administrator to 

ensure that this is ok with local policies -- see if they did any 

special setup for Open MPI, etc.

 

But yes, we made a bunch of COMM_SPAWN improvements since the 1.2 

series.

 

--

Jeff Squyres

 

 

From: Blesson Varghese [mailto:hx019035@reading.ac.uk] 
Sent: 01 October 2009 12:01
To: 'Open MPI Users'; 'Ralph Castain'
Subject: RE: [OMPI users] MPI_Comm_accept()/connect() errors

 

The following is the information regarding the error. I am running Open MPI 1.2.5 on Ubuntu 4.2.4, kernel version 2.6.24

 

I ran the server program as mpirun –np 1 server. This program gave me the output port as 0.1.0:2000. I used this port name value as the command line argument for the client program: mpirun –np 1 client 0.1.1:2000.

 

- The output of the "ompi_info --all" is attached with the email

- PATH Variable: /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/maui/bin/:

- LD_LIBRARY_PATH variable was empty

- The following is the output of ifconfig on hpcc00 from where the error has been generated:

eth0      Link encap:Ethernet  HWaddr 00:12:3f:4c:2d:78

          inet addr:134.225.200.100  Bcast:134.225.200.255  Mask:255.255.255.0

          inet6 addr: fe80::212:3fff:fe4c:2d78/64 Scope:Link

          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

          RX packets:15912728 errors:0 dropped:0 overruns:0 frame:0

          TX packets:15312376 errors:0 dropped:0 overruns:0 carrier:0

          collisions:0 txqueuelen:1000

          RX bytes:2951880321 (2.7 GB)  TX bytes:2788249498 (2.5 GB)

          Interrupt:16

 

lo        Link encap:Local Loopback

          inet addr:127.0.0.1  Mask:255.0.0.0

          inet6 addr: ::1/128 Scope:Host

          UP LOOPBACK RUNNING  MTU:16436  Metric:1

          RX packets:3507489 errors:0 dropped:0 overruns:0 frame:0

          TX packets:3507489 errors:0 dropped:0 overruns:0 carrier:0

          collisions:0 txqueuelen:0

          RX bytes:1794266658 (1.6 GB)  TX bytes:1794266658 (1.6 GB)

 

Regards,

Blesson.

 

From: users-bounces@open-mpi.org [mailto:users-bounces@open-mpi.org] On Behalf Of Ralph Castain
Sent: 29 September 2009 23:59
To: Open MPI Users
Subject: Re: [OMPI users] MPI_Comm_accept()/connect() errors

 

I will ask the obvious - what version of Open MPI are you running? In what environment? What was your command line?

 

:-)

 

On Sep 29, 2009, at 3:50 PM, Blesson Varghese wrote:

 

Hi,

 

I have been trying to execute the server.c and client.c program provided in http://www.mpi-forum.org/docs/mpi21-report/node213.htm#Node213, using accept() and connect() function in MPI. However, the following errors are generated.

 

[hpcc00:16522] *** An error occurred in MPI_Comm_connect

[hpcc00:16522] *** on communicator MPI_COMM_WORLD

[hpcc00:16522] *** MPI_ERR_INTERN: internal error

[hpcc00:16522] *** MPI_ERRORS_ARE_FATAL (goodbye)

 

Could anybody please help me?

 

Many thanks,
Blesson.

_______________________________________________
users mailing list
users@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/users

 

<outputompi.txt>