AMBER Archive (2008)

Subject: RE: AMBER: AMBER Parallel test error

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Tue May 06 2008 - 10:25:46 CDT

Hi Jitrayut,

Problems loading shared libraries normally occur when the environment a job
is run in is different from the environment it was compiled in. For example
on clusters people often compile on a login node but when they submit via a
queuing system it goes to a compute node that is set up differently or does
not inherit the environment variables correctly.

liborte.so.0 is part of openmpi and so you probably don’t have the paths set
correctly by default. I should probably try openmpi myself one of these days
so I understand all of it quirks but I’ve never really had the inclination
since everything else always seems to work fine.

Anyway, try the following.

First find where liborte.so.0 is:

'locate liborte.so.0' may work depending on your OS.

You could also try:

'find ./ -name liborte.so.0'

Or alternatively just manually look for it. I suspect this is in
/usr/local/openmpi/lib or something similar based on where you installed
openmpi.

You then need to add this to the LD_LIBRARY_PATH. I suggest you do this
globally by editing /etc/bashrc (if your shell is bash):

export LD_LIBRARY_PATH=/path/to/openmpi/libdirectory:$LD_LIBRARY_PATH

Then logout, log back in and type 'echo $LD_LIBRARY_PATH' to check it
sticks.

Then try running the mpi job again.

Good luck,
Ross

/\
\/
|\oss Walker

| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

From: owner-amber_at_scripps.edu [mailto:owner-amber_at_scripps.edu] On Behalf Of
jitrayut jitonnom
Sent: Tuesday, May 06, 2008 6:39 AM
To: amber_at_scripps.edu
Subject: Re: AMBER: AMBER Parallel test error

Dear Ross.
 
I am sorry for the early reply. I have checked it again but it 's still
found the problem but not the localhost problem any more. Here is the error
i found when i uses make test.parallel
 
[root_at_chemoinfo test]# make test.parallel
export TESTsander=/usr/local/src/amber9/exe/sander.MPI; make
test.sander.BASIC
make[1]: Entering directory `/usr/local/src/amber9/test'
cd dmp; ./Run.dmp
/usr/local/src/amber9/exe/sander.MPI: error while loading shared libraries:
liborte.so.0: cannot open shared object file: No such file or directory
  ./Run.dmp:  Program error
make[1]: *** [test.sander.BASIC] Error 1
make[1]: Leaving directory `/usr/local/src/amber9/test'
make: *** [test.sander.BASIC.MPI] Error 2
 
To solve with this problem is difficult for me since I am new for unix
installation and I need some learning. So, if anyone can give me some advice
in practical way. It would help me very much.
 
Thanks.
 
Jitrayut

 
2008/5/6, jitrayut jitonnom <jitrayut.018_at_gmail.com>:
Dear Ross Walker,
 
Thank you for you suggestion. As you mention, I forgot to add the line
127.0.0.1 localhost at /etc/host so that why i got error. Besides, when
i try sander.MPI it show a missing library file liborte.so.0 as follow:
 
sander.MPI: error while loading shared libraries: liborte.so.0: cannot open
shared object file: No such file or directory
 
how can I remove this error ?
 
Thanks.
 
Jitrayut

 
2008/5/6, Ross Walker <ross_at_rosswalker.co.uk>:
Hi Jitrayut,
 
This is a problem with your MPI implementation. You should check the
documentation for the version of MPI you are using. If someone manages your
machine or queuing system then check with them to see what specific
requirements there are for running parallel jobs on your system.
 
You should also try some of the simple mpi examples that came with the mpi
implementation you are using since if you can't get these to run there will
be no chance of getting amber to run in parallel.
 
Note from the error message you may need to check that the line:
 
127.0.0.1       localhost.localdomain   localhost
 
Exists in /etc/hosts
 
Although you should also carefully check the mpi documentation to make sure
there is nothing else you need to add.
 
Good luck,
Ross
 
From: owner-amber_at_scripps.edu [mailto:owner-amber_at_scripps.edu] On Behalf Of
jitrayut jitonnom
Sent: Monday, May 05, 2008 11:38 AM
To: amber_at_scripps.edu
Subject: AMBER: AMBER Parallel test error
 
Dear Amber users,
 
I can't complete the parallel test with the command and errors belows, So,
what is my problem and how to solve this problem ?
 

[root_at_chemoinfo test]# make test.parallel
export TESTsander=/usr/local/src/amber9/exe/sander.MPI; make
test.sander.BASIC
make[1]: Entering directory `/usr/local/src/amber9/test'
cd dmp; ./Run.dmp
This test not set up for parallel
 cannot run in parallel with #residues < #pes
cd adenine; ./Run.adenine
This test not set up for parallel
 cannot run in parallel with #residues < #pes
==============================================================
cd cytosine; ./Run.cytosine
ssh: localhost: Name or service not known
[chemoinfo.science.cmu.ac.th:13300] ERROR: A daemon on node localhost failed
to start as expected.
[chemoinfo.science.cmu.ac.th:13300] ERROR: There may be more information
available from
[chemoinfo.science.cmu.ac.th:13300] ERROR: the remote shell (see above).
[chemoinfo.science.cmu.ac.th:13300] ERROR: The daemon exited unexpectedly
with status 255.
  ./Run.cytosine:  Program error
make[1]: *** [test.sander.BASIC] Error 1
 
Jitrayut

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu