See these blog entries for information about Open MPI's tree-based launching:
http://blogs.cisco.com/performance/tree-based-launch-in-open-mpi http://blogs.cisco.com/performance/tree-based-launch-in-open-mpi-part-2 > On May 3, 2015, at 8:52 PM, XingFENG <xingf...@cse.unsw.edu.au> wrote: > > Thanks very much Ralph. It works fine now. > > Just some more questions to make it clearer. > > 1. Your respond confuses me. Especially for this one, The rsh launch is > typically done in a tree topology, and so you get ssh from one backend node > to another. > > 2. By adding -mca plm_rsh_no_tree_spawn 1, it seems that daemon process is > disabled because I could not find processes orted and sshd any more. Is rsh > used instead here? > > > > On Mon, May 4, 2015 at 12:54 AM, Ralph Castain <r...@open-mpi.org> wrote: > >> On May 3, 2015, at 5:54 AM, XingFENG <xingf...@cse.unsw.edu.au> wrote: >> >> By changing the order of hosts, the program now can run on 6 machines with 2 >> accounts. >> >> But it is very weird. Why the order of hosts would matter? Could any one >> refer something about how openmpi work, please? I have searched on the >> Internet but nothing related is found. > > If you add —mca plm_rsh_no_tree_spawn 1 to your cmd line, you’ll probably be > okay. The rsh launch is typically done in a tree topology, and so you get ssh > from one backend node to another. If you don’t have password-less ssh from > every machine to each other machine, then you can get the behavior you > describe. > > Adding that parameter will cause all the ssh commands to come directly from > mpirun. > > >> >> >> On Sun, May 3, 2015 at 10:03 PM, XingFENG <xingf...@cse.unsw.edu.au> wrote: >> I got some problems while running MPI programs with different account. >> I have done the following things. >> >> 1. I have set password-less connection between these machines. >> 2. I have specified user names in host file. >> 3. I have successfully run the program on two machines with two different >> accounts. >> >> But when I tried on other machines with the two accounts, openmpi stuck at >> the very beginning. There is no error information. It seems openmpi falls in >> some infinite loop. Any help would be greatly appreciated. >> >> On Wed, Apr 15, 2015 at 6:04 AM, Jeff Squyres (jsquyres) >> <jsquy...@cisco.com> wrote: >> You can also specify per-machine usernames in $HOME/.ssh/config. >> >> See ssh_config(5). >> >> >> >> > On Apr 13, 2015, at 9:19 PM, Ralph Castain <r...@open-mpi.org> wrote: >> > >> > >> >> On Apr 13, 2015, at 5:47 PM, XingFENG <xingf...@cse.unsw.edu.au> wrote: >> >> >> >> Thanks for all who joined the discussion. >> >> Yes, I meant to run the job under both accounts at the same time. >> >> >> >> Hi, Ralph, >> >> >> >> I have searched about specifying username in hostfile and got no related >> >> page. Could you give a sample of specifying username in hostfile, please? >> > >> > you just provide the host name as “username@host” >> > >> >> Another problem is do you also specify password in hostfile? >> > >> > >> > Good heavens, no! It has to be password-less ssh. Frankly, I doubt this >> > will work as I suspect ssh isn’t going to let you do it. >> > >> > >> >> >> >> On Tue, Apr 14, 2015 at 12:05 AM, Ralph Castain <r...@open-mpi.org> wrote: >> >> >> >>> On Apr 13, 2015, at 6:56 AM, Maxime Boissonneault >> >>> <maxime.boissonnea...@calculquebec.ca> wrote: >> >>> >> >>> Le 2015-04-13 09:54, Ralph Castain a écrit : >> >>>>> On Apr 13, 2015, at 6:52 AM, Maxime Boissonneault >> >>>>> <maxime.boissonnea...@calculquebec.ca> wrote: >> >>>>> >> >>>>> Just out of curiosity... how will OpenMPI start processes under >> >>>>> different accounts ? Through SSH while specifying different user names >> >>>>> ? >> >>>>> I am assuming that no resource manager or scheduler will allow this. >> >>>> I’m assuming he just plans to run the entire job as the other user. >> >>>> Essentially, it would be the same as if his friend ran the job for him. >> >>> From this comment : >> >>> >> >>> My problem is that my account is limited to use 4 machines (I need more >> >>> machines to process data). I can borrow my friend's account and thus >> >>> have access to another 4 machines but I am not sure whether it works. >> >>> >> >>> I assumed that he wants to run the job under _both_ accounts at the same >> >>> time. >> >> >> >> I honestly don’t know if that would work. We do provide a way in the >> >> hostfile for someone to specify the username for a remote host, but you’d >> >> have to get the ssh system to accept it. Assuming the other username also >> >> has your host in its authorized keys, then I suppose it might work - I’ve >> >> never tried it myself. >> >> >> >> >> >>> >> >>> >> >>>>> My recommendation would be to contact your sysadmin and ask for an >> >>>>> exception instead of going through with this insanity (forgive the >> >>>>> judgement here). >> >>>> Agreed! >> >>>> >> >>>>> Maxime >> >>>>> >> >>>>> Le 2015-04-13 09:47, Ralph Castain a écrit : >> >>>>>> Let’s hope you sys admin doesn’t find out about it - they tend to >> >>>>>> take a dim view of sharing accounts! >> >>>>>> >> >>>>>> So long as the path and library path are set correctly, we won’t care. >> >>>>>> >> >>>>>> >> >>>>>>> On Apr 12, 2015, at 10:33 PM, XingFENG <xingf...@cse.unsw.edu.au> >> >>>>>>> wrote: >> >>>>>>> >> >>>>>>> Hi all, >> >>>>>>> >> >>>>>>> I am wondering if it is possible that MPI programs can be run on >> >>>>>>> machines with different account? >> >>>>>>> >> >>>>>>> I am doing experiments with some MPI programs on a cluster. My >> >>>>>>> problem is that my account is limited to use 4 machines (I need more >> >>>>>>> machines to process data). I can borrow my friend's account and thus >> >>>>>>> have access to another 4 machines but I am not sure whether it works. >> >>>>>>> >> >>>>>>> -- >> >>>>>>> Best Regards. >> >>>>>>> --- >> >>>>>>> Xing FENG >> >>>>>>> PhD Candidate >> >>>>>>> Database Research Group >> >>>>>>> >> >>>>>>> School of Computer Science and Engineering >> >>>>>>> University of New South Wales >> >>>>>>> NSW 2052, Sydney >> >>>>>>> >> >>>>>>> Phone: (+61) 413 857 288 >> >>>>>>> _______________________________________________ >> >>>>>>> users mailing list >> >>>>>>> us...@open-mpi.org >> >>>>>>> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> >>>>>>> Link to this post: >> >>>>>>> http://www.open-mpi.org/community/lists/users/2015/04/26687.php >> >>>>>> _______________________________________________ >> >>>>>> users mailing list >> >>>>>> us...@open-mpi.org >> >>>>>> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> >>>>>> Link to this post: >> >>>>>> http://www.open-mpi.org/community/lists/users/2015/04/26690.php >> >>>>> >> >>>>> -- >> >>>>> --------------------------------- >> >>>>> Maxime Boissonneault >> >>>>> Analyste de calcul - Calcul Québec, Université Laval >> >>>>> Ph. D. en physique >> >>>>> >> >>>>> _______________________________________________ >> >>>>> users mailing list >> >>>>> us...@open-mpi.org >> >>>>> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> >>>>> Link to this post: >> >>>>> http://www.open-mpi.org/community/lists/users/2015/04/26691.php >> >>>> _______________________________________________ >> >>>> users mailing list >> >>>> us...@open-mpi.org >> >>>> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> >>>> Link to this post: >> >>>> http://www.open-mpi.org/community/lists/users/2015/04/26692.php >> >>> >> >>> >> >>> -- >> >>> --------------------------------- >> >>> Maxime Boissonneault >> >>> Analyste de calcul - Calcul Québec, Université Laval >> >>> Ph. D. en physique >> >>> >> >>> _______________________________________________ >> >>> users mailing list >> >>> us...@open-mpi.org >> >>> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> >>> Link to this post: >> >>> http://www.open-mpi.org/community/lists/users/2015/04/26693.php >> >> >> >> >> >> _______________________________________________ >> >> users mailing list >> >> us...@open-mpi.org >> >> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> >> Link to this post: >> >> http://www.open-mpi.org/community/lists/users/2015/04/26694.php >> >> >> >> >> >> >> >> -- >> >> Best Regards. >> >> --- >> >> Xing FENG >> >> PhD Candidate >> >> Database Research Group >> >> >> >> School of Computer Science and Engineering >> >> University of New South Wales >> >> NSW 2052, Sydney >> >> >> >> Phone: (+61) 413 857 288 >> >> _______________________________________________ >> >> users mailing list >> >> us...@open-mpi.org >> >> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> >> Link to this post: >> >> http://www.open-mpi.org/community/lists/users/2015/04/26707.php >> > >> > _______________________________________________ >> > users mailing list >> > us...@open-mpi.org >> > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> > Link to this post: >> > http://www.open-mpi.org/community/lists/users/2015/04/26708.php >> >> >> -- >> Jeff Squyres >> jsquy...@cisco.com >> For corporate legal information go to: >> http://www.cisco.com/web/about/doing_business/legal/cri/ >> >> _______________________________________________ >> users mailing list >> us...@open-mpi.org >> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> Link to this post: >> http://www.open-mpi.org/community/lists/users/2015/04/26729.php >> >> >> >> -- >> Best Regards. >> --- >> Xing FENG >> PhD Candidate >> Database Research Group >> >> School of Computer Science and Engineering >> University of New South Wales >> NSW 2052, Sydney >> >> Phone: (+61) 413 857 288 >> >> >> >> -- >> Best Regards. >> --- >> Xing FENG >> PhD Candidate >> Database Research Group >> >> School of Computer Science and Engineering >> University of New South Wales >> NSW 2052, Sydney >> >> Phone: (+61) 413 857 288 >> _______________________________________________ >> users mailing list >> us...@open-mpi.org >> Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users >> Link to this post: >> http://www.open-mpi.org/community/lists/users/2015/05/26831.php > > > _______________________________________________ > users mailing list > us...@open-mpi.org > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users > Link to this post: > http://www.open-mpi.org/community/lists/users/2015/05/26832.php > > > > -- > Best Regards. > --- > Xing FENG > PhD Candidate > Database Research Group > > School of Computer Science and Engineering > University of New South Wales > NSW 2052, Sydney > > Phone: (+61) 413 857 288 > _______________________________________________ > users mailing list > us...@open-mpi.org > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users > Link to this post: > http://www.open-mpi.org/community/lists/users/2015/05/26833.php -- Jeff Squyres jsquy...@cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/
