CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (https://www.cfd-online.com/Forums/openfoam-solving/)
-   -   connect failed with errno%3d113 (https://www.cfd-online.com/Forums/openfoam-solving/58395-connect-failed-errno-3d113.html)

wedsall September 3, 2008 10:36

Hello board, Thanks in adv
 
Hello board,

Thanks in advance to anyone who is able to assist us with this problem.
We are running OpenFoam 1.4.1 on a suse 10.3 cluster. Half of our nodes have infiniband installed, half do not. This case appears to work on the non-infiniband nodes but not on the infiniband nodes. Here is the point at which mpirun is erroring out.

[95] Nprocs : 160
[node024][0,1,104][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
[node020][0,1,72][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
[node026][0,1,120][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
[node028][0,1,136][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
[node009][0,1,8][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
[node022][0,1,88][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
[node014][0,1,32][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
[node030][0,1,152][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113

Does this look familiar to anyone? ANy help will be greatly appreciated.

rangesh November 3, 2008 01:06

hi William, I m a
 
hi William,
I m also facing same problem while running the openFoam 1.5. If u have the solution for that please help me. my emalid is rangeshgupta@gmail.com.
It is showing error as :

[n23][0,1,1][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] [n22][0,1,0][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
[n28][0,1,2][btl_tcp_endpoint.c:572:mca_btl_tcp_endpoint_comple te_connect] connect() failed with errno=113
connect() failed with errno=113

Plz help me on this matter.

thanks

mattijs November 3, 2008 05:17

We had similar messages when w
 
We had similar messages when we had two ethernet cards per node and had to use the 'mca' parameters to have it use eth0 only:

mpirun --mca btl_tcp_if_exclude lo,eth1 --hostfile {hostfile} -np 4 ...

Don't know about infiniband but you might have to do something similar. See the openmpi documentation.


All times are GMT -4. The time now is 23:51.