- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Dear all I made a cluster with 2 nodes.
I installed openmpi with intel compilers and shared it via nfs and also the intel compilers are shared via nfs on each nodes.
when I try to run some applications on each nodes i found:
/opt/openmpi-1.6.5/bin/orted: error while loading shared libraries: libimf.so: cannot open shared object file: No such file or directory
--------------------------------------------------------------------------
A daemon (pid 18804) died unexpectedly with status 127 while attempting
to launch so we are aborting.
There may be more information reported by the environment (see above).
This may be because the daemon was unable to find all the needed shared
libraries on the remote node. You may set your LD_LIBRARY_PATH to have the
location of the shared libraries on the remote nodes and this will
automatically be forwarded to the remote nodes.
So I export in the .bash_profile of each nodes
export LD_LIBRARY_PATH=/opt/intel/lib/intel64:$LD_LIBRARY_PATH
source /opt/intel/bin/compilervars.sh intel64
export PATH=/opt/openmpi-1.6.5/bin:$PATH
export LD_LIBRARY_PATH=/opt/openmpi-1.6.5/lib:$LD_LIBRARY_PATH
But the problem is still the same I am not able to loading shared libraries : liibmf.so
Please can you help me to fix the problem?
Many thanks
lorenzo
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Sorry I'am running under Ubuntu 14.04.5 LTS on each nodes
Thanks again
lorenzo
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
You seem to be on the correct track. You have a bit of duplication in what you added to the .bash_profile. The source of the compilervars will establish the necessary environment variable settings including LD_LIBRARY_PATH so the initial export you added above that line would not be needed.
This FAQ contains exactly (Q #12) the scenario you are facing.
You need to try to determine whether the .bash_profile is invoked when your MPI job runs. Maybe you can drop an echo commend in there to see. You might echo the value of LD_LIBRARY_PATH after the commands to help verify its value.
Have you also confirmed you can run on either node interactively using a simple non-MPI hello_world Fortran program?
Sorry I don't have much direct advice on resolving this but again you seem on the correct path.
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page