From: Eric George <ergeorge@worldnet.att.net>
Newsgroups: comp.parallel.pvm
Subject: Re: trying to use PVM: errors?!?!
Date: 12 Mar 1999 19:38:07 GMT
Organization: AT&T WorldNet Services
Message-Id: <36E95EED.59D48CAA@worldnet.att.net>
References: <36E69038.EFBF5469@postech.ac.kr>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Xref: ukc comp.parallel.pvm:8107


It sounds like pvm can't find the slave executables.  Did you copy the
slave executables to the remote machine?  PVM looks for the executables
for the spawned files in $PVM_ROOT/bin/LINUX  (for linux architectures).

Eric

Rob Lahaye wrote:

> Hi all,
>
> I've installed PVM3.4beta7 sort of successfully in my
> home directories of two PC's both running Linux Redhat5.2.
>
> I can set up two hosts within the pvm console.
> surfion4: local Linux PC
> surfion1: remote Linux PC
>
> =============================================================
>
> [LINUX]$ pvm
> pvm> add surfion1.postech.ac.kr
> 1 successful
>                     HOST     DTID
>   surfion1.postech.ac.kr    80000
> pvm> conf
> 2 hosts, 1 data format
>                     HOST     DTID     ARCH   SPEED       DSIG
>   surfion4.postech.ac.kr    40000    LINUX    1000 0x00408841
>   surfion1.postech.ac.kr    80000    LINUX    1000 0x00408841
> pvm> quit
>
> Console: exit handler called
> pvmd still running.
> [LINUX]$ master1
> Spawning 6 worker tasks ...
>  Trouble spawning slaves. Aborting. Error codes are:
> TID 3 -7
> TID 4 -7
> TID 5 -7
> [LINUX]$
>
> ===============================================================
>
> I've put some additional debug print statements in master1.c
> and I get the following output during runtime:
>
> nproc = 6
> nhost = 2
> numt = 3
>
> tids[0] = 262147
> tids[1] = 262148
> tids[2] = 262149
> tids[3] = -7
> tids[4] = -7
> tids[5] = -7
>
> When I delete the remote PC and use only my local PC,
> everything goes fine.
> What else then "add" is necessary to connect to the
> remote Linux PC?
>
> Can anybody tell me what could be the problem?
>
> Thanks,
>
> Rob.