Running in Parallel on cluster

NewFoamer · November 2, 2010, 09:15

Hello,

I'm trying to run a cache on my school cluster. Currently I'm issuing following command:

Code:

mpirun --hostfile system/machines -np 6 pimpleFoamScalar -parallel

With my hostfile:

Code:

XXX@top.nbi.dk slots=2 max-slots=2
XXX@charm.nbi.dk slots=2 max-slots=2
XXX@bottom.nbi.dk slots=2 max-slots=2

Which gives me following error:

Code:

[top.nbi.dk:18796] btl: tcp: attempting to connect() to address 130.225.212.8 on port 260
[top.nbi.dk][[19986,1],0][btl_tcp_endpoint.c:638:mca_btl_tcp_endpoint_complete_connect] connect() to 130.225.212.8 failed: No route to host (113)

From the google-searched I've done, I believe I have to properly set the --mca to something, but I've thus far been unsuccessful in doing so properly. I'm new to OpenMPI, and do not have a lot of experience with network interfaces etc.

Any help or clues as how to debug the problem would be appreciated!
Thank you.

falcao · November 2, 2010, 10:57

There is a OpenFoam cluster receipe, in the second page, in this same forum.

http://www.cfd-online.com/Forums/ope...am-solved.html

The translation to English is immediately below.

NewFoamer · November 2, 2010, 11:04

This does not solve my problem - The machines share the same network drive and I've decomposed, made the hostfile etc. correctly.

r08n · November 3, 2010, 16:20

[top.nbi.dk][[19986,1],0][btl_tcp_endpoint.c:638:mca_btl_tcp_endpoint_comple te_connect] connect() to 130.225.212.8 failed: No route to host (113)

This looks like a network problem (misconfiguration?). First of all, check if you can
connect from one node of the cluster to any another node; e.g., like this:
telnet 130.225.212.8 260; if you get this message "no route to host", check the network configuration.

November 2, 2010, 09:15	Running in Parallel on cluster	#1
NewFoamer New Member Join Date: Aug 2010 Posts: 14 Rep Power: 15	Hello, I'm trying to run a cache on my school cluster. Currently I'm issuing following command: Code: mpirun --hostfile system/machines -np 6 pimpleFoamScalar -parallel With my hostfile: Code: XXX@top.nbi.dk slots=2 max-slots=2 XXX@charm.nbi.dk slots=2 max-slots=2 XXX@bottom.nbi.dk slots=2 max-slots=2 Which gives me following error: Code: [top.nbi.dk:18796] btl: tcp: attempting to connect() to address 130.225.212.8 on port 260 [top.nbi.dk][[19986,1],0][btl_tcp_endpoint.c:638:mca_btl_tcp_endpoint_complete_connect] connect() to 130.225.212.8 failed: No route to host (113) From the google-searched I've done, I believe I have to properly set the --mca to something, but I've thus far been unsuccessful in doing so properly. I'm new to OpenMPI, and do not have a lot of experience with network interfaces etc. Any help or clues as how to debug the problem would be appreciated! Thank you.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
problem of running parallel Fluent on linux cluster	ivanbuz	FLUENT	15	September 23, 2017 19:12
running OpenFoam in parallel	vishwa	OpenFOAM Running, Solving & CFD	22	August 2, 2015 08:53
Script to Run Parallel Jobs in Rocks Cluster	asaha	OpenFOAM Running, Solving & CFD	12	July 4, 2012 22:51
Statically Compiling OpenFOAM Issues	herzfeldd	OpenFOAM Installation	21	January 6, 2009 09:38
Kubuntu uses dash breaks All scripts in tutorials	platopus	OpenFOAM Bugs	8	April 15, 2008 07:52

November 2, 2010, 10:57		#2
falcao New Member Join Date: Feb 2010 Posts: 13 Rep Power: 16	There is a OpenFoam cluster receipe, in the second page, in this same forum. http://www.cfd-online.com/Forums/ope...am-solved.html The translation to English is immediately below.

November 2, 2010, 11:04		#3
NewFoamer New Member Join Date: Aug 2010 Posts: 14 Rep Power: 15	This does not solve my problem - The machines share the same network drive and I've decomposed, made the hostfile etc. correctly.

November 3, 2010, 16:20		#4
r08n Member Robertas N. Join Date: Mar 2009 Location: Kaunas, Lithuania Posts: 53 Rep Power: 17	[top.nbi.dk][[19986,1],0][btl_tcp_endpoint.c:638:mca_btl_tcp_endpoint_comple te_connect] connect() to 130.225.212.8 failed: No route to host (113) This looks like a network problem (misconfiguration?). First of all, check if you can connect from one node of the cluster to any another node; e.g., like this: telnet 130.225.212.8 260; if you get this message "no route to host", check the network configuration.