Open MPI logo

Open MPI User's Mailing List Archives

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

Subject: Re: [OMPI users] client-server example
From: John Chludzinski (john.chludzinski_at_[hidden])
Date: 2013-04-13 19:54:40


After I replaced ";" with "\;" in the server name I got passed the
ABORT problem. Now the client and server deadlock until I finally get
(on the client side):

mpirun noticed that the job aborted, but has no info as to the process
that caused that situation.
--------------------------------------------------------------------------
[jski:02429] [[59675,0],0] -> [[59187,0],0] (node: jski) oob-tcp:
Number of attempts to create TCP connection has been exceeded. Cannot
communicate with peer.

On Sat, Apr 13, 2013 at 7:24 PM, John Chludzinski
<john.chludzinski_at_[hidden]> wrote:
> Sorry: The previous post was intended for another group, ignore it.
>
> With regards to the client-server problem:
>
> $ mpirun -n 1 client
> 3878879232.0;tcp://192.168.1.4:37625+3878879233.0;tcp://192.168.1.4:38945:300
>
> [jski:01882] [[59199,1],0] ORTE_ERROR_LOG: Not found in file
> dpm_orte.c at line 158
> [jski:1882] *** An error occurred in MPI_Comm_connect
> [jski:1882] *** on communicator MPI_COMM_WORLD
> [jski:1882] *** MPI_ERR_INTERN: internal error
> [jski:1882] *** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
> --------------------------------------------------------------------------
> mpirun has exited due to process rank 0 with PID 1882 on
> node jski exiting improperly. There are two reasons this could occur:
>
> 1. this process did not call "init" before exiting, but others in
> the job did. This can cause a job to hang indefinitely while it waits
> for all processes to call "init". By rule, if one process calls "init",
> then ALL processes must call "init" prior to termination.
>
> 2. this process called "init", but exited without calling "finalize".
> By rule, all processes that call "init" MUST call "finalize" prior to
> exiting or it will be considered an "abnormal termination"
>
> On Sat, Apr 13, 2013 at 7:16 PM, John Chludzinski
> <john.chludzinski_at_[hidden]> wrote:
>> After I "source mpi.ksk", PATH is unchanged but LD_LIBRARY_PATH is there:
>>
>> $ print $LD_LIBRARY_PATH
>> /usr/lib64/openmpi/lib/
>>
>> Why does PATH loose its change?
>>
>> ---John
>>
>>
>> On Sat, Apr 13, 2013 at 12:55 PM, Ralph Castain <rhc_at_[hidden]> wrote:
>>> You need to pass in the port info that the server printed - just copy/paste the line below "server available at".
>>>
>>> On Apr 12, 2013, at 10:58 PM, John Chludzinski <john.chludzinski_at_[hidden]> wrote:
>>>
>>>> Found the following client-server example (code) on
>>>> http://www.mpi-forum.org and I'm trying to get it to work. Not sure
>>>> what argv[1] should be for the client? The output from the server
>>>> side is:
>>>>
>>>> server available at
>>>> 4094230528.0;tcp://192.168.1.4:55803+4094230529.0;tcp://192.168.1.4:51618:300
>>>>
>>>>
>>>> // SERVER
>>>> #include <stdio.h>
>>>> #include <error.h>
>>>> #include <errno.h>
>>>> #include "mpi.h"
>>>>
>>>> #define MAX_DATA 100
>>>> #define FATAL 1
>>>>
>>>> int main( int argc, char **argv )
>>>> {
>>>> MPI_Comm client;
>>>> MPI_Status status;
>>>> char port_name[MPI_MAX_PORT_NAME];
>>>> double buf[MAX_DATA];
>>>> int size, again;
>>>>
>>>> MPI_Init( &argc, &argv );
>>>> MPI_Comm_size(MPI_COMM_WORLD, &size);
>>>> if (size != 1) error(FATAL, errno, "Server too big");
>>>> MPI_Open_port(MPI_INFO_NULL, port_name);
>>>> printf("server available at %s\n",port_name);
>>>>
>>>> while (1)
>>>> {
>>>> MPI_Comm_accept( port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &client );
>>>> again = 1;
>>>>
>>>> while (again)
>>>> {
>>>> MPI_Recv( buf, MAX_DATA, MPI_DOUBLE, MPI_ANY_SOURCE,
>>>> MPI_ANY_TAG, client, &status );
>>>>
>>>> switch (status.MPI_TAG)
>>>> {
>>>> case 0: MPI_Comm_free( &client );
>>>> MPI_Close_port(port_name);
>>>> MPI_Finalize();
>>>> return 0;
>>>> case 1: MPI_Comm_disconnect( &client );
>>>> again = 0;
>>>> break;
>>>> case 2: /* do something */
>>>> fprintf( stderr, "Do something ...\n" );
>>>> default:
>>>> /* Unexpected message type */
>>>> MPI_Abort( MPI_COMM_WORLD, 1 );
>>>> }
>>>> }
>>>> }
>>>> }
>>>>
>>>> //CLIENT
>>>> #include <string.h>
>>>> #include "mpi.h"
>>>>
>>>> #define MAX_DATA 100
>>>>
>>>> int main( int argc, char **argv )
>>>> {
>>>> MPI_Comm server;
>>>> double buf[MAX_DATA];
>>>> char port_name[MPI_MAX_PORT_NAME];
>>>> int done = 0, tag, n, CNT=0;
>>>>
>>>> MPI_Init( &argc, &argv );
>>>> strcpy(port_name, argv[1] ); /* assume server's name is cmd-line arg */
>>>>
>>>> MPI_Comm_connect( port_name, MPI_INFO_NULL, 0, MPI_COMM_WORLD, &server );
>>>>
>>>> n = MAX_DATA;
>>>>
>>>> while (!done)
>>>> {
>>>> tag = 2; /* Action to perform */
>>>> if ( CNT == 5 ) { tag = 0; done = 1; }
>>>> MPI_Send( buf, n, MPI_DOUBLE, 0, tag, server );
>>>> CNT++;
>>>> /* etc */
>>>> }
>>>>
>>>> MPI_Send( buf, 0, MPI_DOUBLE, 0, 1, server );
>>>> MPI_Comm_disconnect( &server );
>>>> MPI_Finalize();
>>>>
>>>> return 0;
>>>> }
>>>> _______________________________________________
>>>> users mailing list
>>>> users_at_[hidden]
>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>
>>>
>>> _______________________________________________
>>> users mailing list
>>> users_at_[hidden]
>>> http://www.open-mpi.org/mailman/listinfo.cgi/users