[Pw_forum] runtime error : [btl_tcp_endpoint.c:655:mca_btl_tcp_endpoint_complete_connect] connect() to 192.168.5.164 failed: Connection refused (111)

[Mike Marchywka]

If I just put that message into google it comes up as something to do witj MPI and
if I read it right it is complaining about port 111 which would be plausible for portmap
or some similar service. Any way, if it is unable to find port 111 did you try to use 
netstat or nmap to see if the expected service is running and observable from the other machines?

If you can find enough key words you can probably find something in your LAN or QE or some supporting servuce
to adjust. 


________________________________
> Date: Wed, 3 Jul 2013 10:20:59 +0530 
> From: dhirendra22121987 at gmail.com 
> To: pw_forum at pwscf.org 
> Subject: Re: [Pw_forum] runtime error : 
> [btl_tcp_endpoint.c:655:mca_btl_tcp_endpoint_complete_connect] 
> connect() to 192.168.5.164 failed: Connection refused (111) 
> 
> Actually I am running pw.x through PBS using qsub. System admin says 
> passwardless ssh is taken care. I can run small parallel programs but 
> when I try to run pw.x it gives above error. 
> 
> 
> On Wed, Jul 3, 2013 at 8:17 AM, 
> <foster362 at gmail.com<mailto:foster362 at gmail.com>> wrote: 
> Can you ssh between nodes (node1 to node2) and server without having to 
> enter a password? 
> 
> On Jul 2, 2013, at 2:52 PM, DHIRENDRA VAIDYA 
> <dhirendra22121987 at gmail.com<mailto:dhirendra22121987 at gmail.com>> 
> wrote: 
> 
> Dear PW_forum community, 
> 
> I am new to Quantum Espresso and I am trying to instal QE on our 
> institute cluster. Compilers are installed in some non-standard 
> directory and care is taken to include paths to variable PATH and 
> LD_LIBRARY_PATH. 
> 
> With this an espresso-5.0.1 is installed with gcc-4.8.1 and 
> openmpi-1.6.5. I tried to produce bandstructure of silicon (given in 
> one of the tutorials on web). By setting #PBS -l nodes=1:ppn=4 I get 
> the bandstructure of silicon, the problem is in when #PBS -l nodes=2(or 
> higher):ppn=2(anything here). The error I get is 
> 
> [compute-0-94.local][[63421,1],2][btl_tcp_endpoint.c:655:mca_btl_tcp_endpoint_complete_connect] 
> connect() to 192.168.5.164 failed: Connection refused (111) 
> [compute-0-94.local][[63421,1],3][btl_tcp_endpoint.c:655:mca_btl_tcp_endpoint_complete_connect] 
> connect() to 192.168.5.164 failed: Connection refused (111) 
> =>> PBS: job killed: walltime 607 exceeded limit 600 
> mpiexec: killing job... 
> 
> 
> Neither I nor our system admin doesnt know about this error. Can 
> someone please tell me what this error is about. There is very little 
> help on web about this error and solution to add --mca [with some 
> options] doesnt seem to work for me. 
> -- 
> Thanks a lot, 
> Dhirendra Vaidya, 
> IIT Bombay 
> _______________________________________________ 
> Pw_forum mailing list 
> Pw_forum at pwscf.org<mailto:Pw_forum at pwscf.org> 
> http://pwscf.org/mailman/listinfo/pw_forum 
> 
> _______________________________________________ 
> Pw_forum mailing list 
> Pw_forum at pwscf.org<mailto:Pw_forum at pwscf.org> 
> http://pwscf.org/mailman/listinfo/pw_forum 
> 
> 
> 
> -- 
> Warm Regards, 
> Dhirendra Vaidya 
> 
> _______________________________________________ Pw_forum mailing list 
> Pw_forum at pwscf.org http://pwscf.org/mailman/listinfo/pw_forum