CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   Problem of cluster (https://www.cfd-online.com/Forums/fluent/90314-problem-cluster.html)

aerodynamics July 6, 2011 20:37

Problem of cluster
 
Hello,
Currently I am trying to run fluent on a cluster of 5 debian computer. The problem is that when I launch fluent in command line with this command :
fluent 3d -t2 -cnf=/home/****/hosts.txt (**** is my usr name) with host.txt the list on my cluster computer, fluent starts and then stop while it is conencting to the other computer.

I think it is due to the fact that the other computer "don't listen" on the network for running in parallel processing. Also cortex is not running.

So do you know how to fix that or what can be the source of the pb.

Thank you in advance.

Galileo July 6, 2011 21:49

Quote:

Originally Posted by aerodynamics (Post 315034)
Hello,
Currently I am trying to run fluent on a cluster of 5 debian computer. The problem is that when I launch fluent in command line with this command :
fluent 3d -t2 -cnf=/home/****/hosts.txt (**** is my usr name) with host.txt the list on my cluster computer, fluent starts and then stop while it is conencting to the other computer.

I think it is due to the fact that the other computer "don't listen" on the network for running in parallel processing. Also cortex is not running.

So do you know how to fix that or what can be the source of the pb.

Thank you in advance.

If the hosts.txt is the journal file, make sure it is in unix format, ie run dos2unix hosts.txt on the command line in your Linux/Unix OS to convert the .txt file and then try again.

Am not sure of what the problem is anyway, just try it if the .txt was a Windows(dos) file. Also make sure you have valid Fluent license on the systems.

Someone with the experience would help if it doesn't work.

aerodynamics July 6, 2011 21:59

I had tryed with different format and it doesn't work. The fact is that if I try on my local machine it is working, even with several processors. And it is not a network pb it has been checked too...

aerodynamics July 10, 2011 22:51

noone has a idea?

Nikolopoulos July 11, 2011 08:53

what interconnection do you use?
SSh or Rsh?

What fluent reports before crashing?

if you cannot find out use the "nohup" command and create a screen.out file


P.S. why did you set -t2 ??? this means 2 cpus. I guess you mean -t5 (or more)


All times are GMT -4. The time now is 08:45.