[Wolf32]

Hi, I've recently installed OpenFOAM 3.0.x and dev on a small cluster (two nodes, each with 2x Xeon 2630 v3 for a total of 32 cores, in a folder shared over NFS) carefully following OpenFOAM Wiki guide and while trying a parallel run with foamJob I've got, with both versions, the same error described in many threads:

orted: Command not found.
--------------------------------------------------------------------------
A daemon (pid ***) died unexpectedly with status 1 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.
--------------------------------------------------------------------------
--------------------------------------------------------------------------
mpirun noticed that the job aborted, but has no info as to the process
that caused that situation.
--------------------------------------------------------------------------

The most meaningful thread is this one. Following the instructions by wyldckat in this post everything worked out well.

On the other hand, while struggling to make it work using just foamJob, I found that simply adding (in a clean installation) the symbolic link:

ln -s /usr/bin/orted $HOME/OpenFOAM/OpenFOAM-dev/bin/orted

makes everything work correctly without problems.

Is this link able to cause problems in some situations? If not, why it's not present in the OpenFOAM installation wiki?
Maybe there is something I don't get, sorry for any stupid question present in this post. I'm writing just to understand better this issue on which I've spent two weeks.

W.

[wyldckat]

Originally Posted by Wolf32

Is this link able to cause problems in some situations? If not, why it's not present in the OpenFOAM installation wiki?
Maybe there is something I don't get, sorry for any stupid question present in this post. I'm writing just to understand better this issue on which I've spent two weeks.

Quick answer: Many thanks for sharing the solution you've found! The symbolic shouldn't be a problem. This specific instruction isn't on the wiki page, because there was no previous report on this

Quick question: Did you do step #5 that is on the wiki page: http://openfoamwiki.net/index.php/In...u#Ubuntu_16.04 ?

Because that step should have solved the issue with orted... although the problem is that the instructions that are on the wiki are usually only tested with builds done on a single virtual machine and subsequent tests are not performed with running in parallel with multiple machines. This is because preparing installations for running in parallel with multiple machines enters into a multi-branching installation instruction system, which depends on the parallellization infrastructure that is being used .
And on the global topic of running OpenFOAM in parallel, the information is still in a planning stage: Notes about running OpenFOAM in parallel

[Wolf32]

Yes, I did step #5 of the guide, but without adding a symbolic link also for orted costantly produced the same error when running in parallel through Infiniband or Ethernet .

I understand my problem can't be handled by OpenFOAM Wiki Installation Guide which is mainly oriented towards installation on a single machine.

On the other hand after some more testing, I confirm that the same symbolic link (as expected) resolves the problem also on Ubuntu 15.10 and Ubuntu 14.04 LTS.

Thanks
Wolf