CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   mpid: No such file or directory [urgent] (http://www.cfd-online.com/Forums/fluent/78418-mpid-no-such-file-directory-urgent.html)

myfluentid July 21, 2010 10:44

mpt_connect: error: connect failed: Connection refused [urgent]
 
Hi everyone,

I've installed Fluent 6.3.26 on Ubuntu 10.04 64bit and now I want to user parallel processing. But I have a problem: After my settings (default interconnection=>ssh) I click on Run button on File->Run menu, the following message appears:

Code:

Host spawning Node 0 on machine "ubuntu" (unix).
/home/Fluent.Inc/fluent6.3.26/bin/fluent -r6.3.26 3d -node -t4 -pethernet -mpi=hp -cnf=/home/hosts.txt -mport 127.0.0.1:127.0.0.1:33873:0
Starting /home/Fluent.Inc/fluent6.3.26/multiport/mpi/lnamd64/hp/bin/mpirun -TCP -f /tmp/fluent-appfile.12875
bash: /home/Fluent.Inc/fluent6.3.26/multiport/mpi/lnamd64/hp/bin/mpid: No such file or directory
mpirun: Warning one more more remote shell commands exited with non-zero status, which may indicate a remote access problem.

and when I try "net interconnection" this message appears:
Code:

Host spawning Node 0 on machine "linlab-desktop" (unix).
/home/Fluent.Inc/fluent6.3.26/bin/fluent -r6.3.26 3d -cl net -node -t4 -pethernet -mpi=net -cnf=/home/hosts.txt -path/home/Fluent.Inc -display :0.0 -mport 127.0.0.1:172.16.124.101:49435:0
Starting /home/Fluent.Inc/fluent6.3.26/lnx86/3d_node/fluent_net.6.3.26 net node -mpiw net -pic ethernet -mport 127.0.0.1:172.16.124.101:49435:0

0: mpt_connect: error: connect failed: Connection refused

0: mpt_establish_connection: error: unable to connect: Invalid argument

0: mpt_connect: error: connect failed: Connection refused

0: mpt_establish_connection: error: unable to connect: Connection refused

0: mpt_connect: error: connect failed: Connection refused

0: mpt_establish_connection: error: unable to connect: Connection refused

0: mpt_connect_to_server: error: cannot establish connection; bye.: Connection refused

I'm confused! I configured ssh with no password. Any more configuration is needed? What should I do in clients?

Thanks in advance

Chris D July 21, 2010 12:08

Do you have fluent installed on all the nodes in your cluster, or have you setup an nfs share?

myfluentid July 22, 2010 07:30

No, I haven't installed on clusters! I thought that fluent will do it itself (by ssh with no pass!). I should do that? How about NFS? Please tell me more. I can't find any tutorial to use parallel fluent, but I'm using fluent on my system successfully.
Any help is appreciated.

Thanks in advance

myfluentid July 22, 2010 07:55

[urgent] please help
 
And I have disabled ufw and apparmor. I installed fluent on clusters and now with default interconnection, error is changed to this:
Code:

mpid: cannot connect to mpirun(127.0.1.1, 34671): Connection refused!
And with net interconnections the same error. I think reason can be one of these:
- Any other service should be installed or started (or something to stop) because of "connection refused"
- Any more configuration for ssh?
why fluent tries to connect 127.0.1.1 instead of 127.0.0.1 ?
what is mpt? it should be installed or configured? it can't even connect !

Would mind help me with fluent and nfs sharing?

Thanks a lot

myfluentid July 22, 2010 08:09

Hi again,
more information:
my /etc/hosts is like this:
Code:

127.0.0.1            localhost
127.0.1.1            linlab-desktop
172.16.123.123  other_computer

And when I change "localhost" and "linlab-desktop", all interconnection errors are:
mpt_connect:error: connect failed: Connection refused.

I have configured ssh with no password and rsh is not installed.

Thanks for any help

myfluentid July 22, 2010 11:56

New errors
 
Hi again,
this is my new error:
Code:

Host spawning Node 0 on machine "my-laptop" (unix).
/home/Fluent.Inc/fluent6.3.26/bin/fluent -r6.3.26 3d -node -t4 -pethernet -mpi=hp -cnf=/home/hosts.txt -mport 127.0.0.1:127.0.0.1:51462:0
Starting /home/Fluent.Inc/fluent6.3.26/multiport/mpi/lnx86/hp/bin/mpirun -TCP -f /tmp/fluent-appfile.3676
mpirun: hostname: linlab-desktop, ip returned: 172.16.124.101
mpirun: hostname: my-laptop, ip returned: 127.0.0.1
mpirun: Can't talk to a remote host with IP 127.0.0.1
mpirun: Change hostname or check IP configuration

What's wrong? anyone can answer? I installed fluent parallel on 2 windows xp, it's working, but in linux I get new errors every time :eek:

this is my hosts.txt:
Code:

linlab-desktop
linlab-desktop
my-laptop
my-laptop


and when I remove "my-laptop" from hosts.txt, error is this:
Code:

Host spawning Node 0 on machine "my-laptop" (unix).
/home/Fluent.Inc/fluent6.3.26/bin/fluent -r6.3.26 3d -node -t4 -pethernet -mpi=hp -cnf=/home/hosts.txt -mport 127.0.0.1:127.0.0.1:33230:0
Starting /home/Fluent.Inc/fluent6.3.26/multiport/mpi/lnx86/hp/bin/mpirun -TCP -f /tmp/fluent-appfile.4204
HP-MPI licensed for execution of Fluent.

0: mpt_connect: error: connect failed: Connection refused

0: mpt_establish_connection: error: unable to connect: Invalid argument

0: mpt_connect: error: connect failed: Connection refused

0: mpt_establish_connection: error: unable to connect: Connection refused

0: mpt_connect: error: connect failed: Connection refused

0: mpt_establish_connection: error: unable to connect: Connection refused

0: mpt_connect_to_server: error: cannot establish connection; bye.: Connection refused
MPI Application rank 0 exited before MPI_Finalize() with status 0

My problem is with ubuntu? all computers should be the same arch and hardware? I have installed fluent on all computers and ssh with no password is working from server to clients. I'm really confused :confused:

Any help is appreciated.

Thanks in advance

Chris D July 22, 2010 13:37

Quote:

Originally Posted by myfluentid (Post 268544)
...ssh with no password is working from server to clients.

I'm not 100% sure, but you might need to be able to ssh both from server to client and from client to server.

ecfd July 22, 2010 17:12

run FLUENT parallel
 
You need to do either of the following:

1. Install FLUENT on the same location on all the involved machines

OR

2. Run FLUENT from a NFS directory, and make sure all the involved machines can access the FLUENT installation with the proper permission


Anyway, I recommend you to use FLUENT 12.1.4, instead of 6.3.26.

myfluentid July 24, 2010 06:56

Hi,

Still the same error:
Code:

0: mpt_connect: error: connect failed: Connection refused
What's wrong? I'm getting crazy about this! Every configuration is ok. Any log file? How can I get more details? Can I change the port it's using? Fluent is working on windows, so what's the problem in linux (the same error in centos and ubuntu!). Running every program in linux is annoying me!!!

Any help is appreciated

majas August 19, 2010 09:20

gambit for linux
 
hi everyone
I want to install gambit on Ubunto system
which version of gambit usable?
do I need to install exceed on it else?
thanks..

fluent-newbie October 22, 2013 17:00

did you ever find a soultion to your problem?


All times are GMT -4. The time now is 11:57.