Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: Re: [OMPI users] [EXTERNAL] Re: open-mpi on Mac OS 10.9 (Mavericks)
From: Meredith, Karl (karl.meredith_at_[hidden])
Date: 2013-12-03 07:44:52


Using the latest nightly snapshot (1.7.4) and only Apple compilers/tools (no macports), I configure/build with the following:

./configure --prefix=/opt/trunk/apple-only-1.7.4 --enable-shared --disable-static --enable-debug --disable-io-romio --enable-contrib-no-build=vt,libtrace --enable-mpirun-prefix-by-default
make all
make install
export PATH=/opt/trunk/apple-only-1.7.4/bin/:$PATH
export LD_LIBRARY_PATH=/opt/trunk/apple-only-1.7.4/lib:$LD_LIBRARY_PATH
export DYLD_LIBRARY_PATH=/opt/trunk/apple-only-1.7.4/lib:$DYLD_LIBRARY_PATH
cd examples
make all
mpirun -v -np 2 ./hello_cxx

Here’s the stack trace for one of the hanging process:

(lldb) bt
* thread #1: tid = 0x57052, 0x00007fff8c991a3a libsystem_kernel.dylib`__semwait_signal + 10, queue = 'com.apple.main-thread, stop reason = signal SIGSTOP
    frame #0: 0x00007fff8c991a3a libsystem_kernel.dylib`__semwait_signal + 10
    frame #1: 0x00007fff8ade4e60 libsystem_c.dylib`nanosleep + 200
    frame #2: 0x0000000100be98e3 libopen-rte.6.dylib`orte_routed_base_register_sync(setup=true) + 2435 at routed_base_fns.c:344
    frame #3: 0x0000000100ecc3a7 mca_routed_binomial.so`init_routes(job=1305542657, ndat=0x0000000000000000) + 2759 at routed_binomial.c:708
    frame #4: 0x0000000100b9e84d libopen-rte.6.dylib`orte_ess_base_app_setup(db_restrict_local=true) + 2109 at ess_base_std_app.c:233
    frame #5: 0x0000000100e3a442 mca_ess_env.so`rte_init + 418 at ess_env_module.c:146
    frame #6: 0x0000000100b59cfe libopen-rte.6.dylib`orte_init(pargc=0x0000000000000000, pargv=0x0000000000000000, flags=32) + 718 at orte_init.c:158
    frame #7: 0x00000001008bd3c8 libmpi.1.dylib`ompi_mpi_init(argc=0, argv=0x0000000000000000, requested=0, provided=0x00007fff5f3cd370) + 616 at ompi_mpi_init.c:451
    frame #8: 0x000000010090b5c3 libmpi.1.dylib`MPI_Init(argc=0x0000000000000000, argv=0x0000000000000000) + 515 at init.c:86
    frame #9: 0x0000000100833a1d hello_cxx`MPI::Init() + 29 at functions_inln.h:128
    frame #10: 0x00000001008332ac hello_cxx`main(argc=1, argv=0x00007fff5f3cd550) + 44 at hello_cxx.cc:18
    frame #11: 0x00007fff8d5df5fd libdyld.dylib`start + 1

Karl

On Dec 2, 2013, at 2:33 PM, Jeff Squyres (jsquyres) <jsquyres_at_[hidden]> wrote:

> Ah -- sorry, I missed this mail before I replied to the other thread (OS X Mail threaded them separately somehow...).
>
> Sorry to ask you to dive deeper, but can you find out where in orte_ess.init() it's failing? orte_ess.init is actually a function pointer; it's a jump-off point into a dlopen'ed plugin.
>
>
> On Nov 25, 2013, at 11:53 AM, "Meredith, Karl" <karl.meredith_at_[hidden]> wrote:
>
>> Digging a little deeper by running the code in the lldb debugger, I found that the stall occurs in a call to init_orte from ompi_mpi_init.c:
>> 356 /* Setup ORTE - note that we are an MPI process */
>> 357 if (ORTE_SUCCESS != (ret = orte_init(NULL, NULL, ORTE_PROC_MPI))) {
>> 358 error = "ompi_mpi_init: orte_init failed";
>> 359 goto error;
>> 360 }
>>
>> The code never returns from orte_init.
>>
>> It gets stuck in orte_ess.init() called from orte_init.c:
>> 126 /* initialize the RTE for this environment */
>> 127 if (ORTE_SUCCESS != (ret = orte_ess.init())) {
>>
>> When I step through this orte_ess_init in the lldb debugger, I actually get some output from the code (no output if not using the debugger and stepping through):
>> --------------------------------------------------------------------------
>> It looks like MPI_INIT failed for some reason; your parallel process is
>> likely to abort. There are many reasons that a parallel process can
>> fail during MPI_INIT; some of which are due to configuration or environment
>> problems. This failure appears to be an internal failure; here's some
>> additional information (which may only be relevant to an Open MPI
>> developer):
>>
>> ompi_mpi_init: orte_init failed
>> --> Returned "Unable to start a daemon on the local node" (-128) instead of "Success" (0)
>>
>>
>>
>> Karl
>>
>>
>>
>> On Nov 25, 2013, at 9:20 AM, Meredith, Karl <karl.meredith_at_[hidden]> wrote:
>>
>>> Here’s the back trace from lldb:
>>> $ )ps -elf | grep hello
>>> 1042653210 45231 45230 4006 0 31 0 2448976 2148 - S+ 0 ttys002 0:00.01 hello_cxx 9:07AM
>>> 1042653210 45232 45230 4006 0 31 0 2457168 2156 - S+ 0 ttys002 0:00.04 hello_cxx 9:07AM
>>>
>>> (meredithk_at_meredithk-mac)-(09:15 AM Mon Nov 25)-(~/tools/openmpi-1.6.5/examples)
>>> $ )lldb -p 45231
>>> Attaching to process with:
>>> process attach -p 45231
>>> Process 45231 stopped
>>> Executable module set to "/Users/meredithk/tools/openmpi-1.6.5/examples/hello_cxx".
>>> Architecture set to: x86_64-apple-macosx.
>>> (lldb) bt
>>> * thread #1: tid = 0x168535, 0x00007fff8c1859aa libsystem_kernel.dylib`select$DARWIN_EXTSN + 10, queue = 'com.apple.main-thread, stop reason = signal SIGSTOP
>>> frame #0: 0x00007fff8c1859aa libsystem_kernel.dylib`select$DARWIN_EXTSN + 10
>>> frame #1: 0x0000000106b73ea0 libmpi.1.dylib`select_dispatch(base=0x00007f84c3c0b430, arg=0x00007f84c3c0b3e0, tv=0x00007fff5924ca70) + 80 at select.c:174
>>> frame #2: 0x0000000106b3eb0f libmpi.1.dylib`opal_event_base_loop(base=0x00007f84c3c0b430, flags=5) + 415 at event.c:838
>>>
>>> Both processors are at this state.
>>>
>>> Here’s the output from otool -L ./hello_cxx:
>>>
>>> $ )otool -L ./hello_cxx
>>> ./hello_cxx:
>>> /Users/meredithk/tools/openmpi/lib/libmpi_cxx.1.dylib (compatibility version 2.0.0, current version 2.2.0)
>>> /Users/meredithk/tools/openmpi/lib/libmpi.1.dylib (compatibility version 2.0.0, current version 2.8.0)
>>> /opt/local/lib/libgcc/libstdc++.6.dylib (compatibility version 7.0.0, current version 7.18.0)
>>> /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1197.1.1)
>>> /opt/local/lib/libgcc/libgcc_s.1.dylib (compatibility version 1.0.0, current version 1.0.0)
>>>
>>>
>>> On Nov 25, 2013, at 9:14 AM, George Bosilca <bosilca_at_[hidden]> wrote:
>>>
>>>> Mac OS X 1.9 dropped support for gdb. Please report the output of lldb instead.
>>>>
>>>> Also, can you run “otool -L ./hello_cxx” and report the output.
>>>>
>>>> Thanks,
>>>> George.
>>>>
>>>>
>>>> On Nov 25, 2013, at 15:09 , Meredith, Karl <karl.meredith_at_[hidden]> wrote:
>>>>
>>>>> I do have DYLD_LIBRARY_PATH set to the same paths as LD_LIBRARY_PATH. This does not resolve the problem. The code still hangs on MPI::Init().
>>>>>
>>>>> Another thing I tried is I recompiled openmpi with the debug flags activated:
>>>>> ./configure --prefix=$HOME/tools/openmpi --enable-debug
>>>>> make
>>>>> make install
>>>>>
>>>>> Then, I attached to the running process using gdb. I tried to do a back trace and see where it was hanging up at, but all I got was this:
>>>>> Attaching to process 45231
>>>>> Reading symbols from /Users/meredithk/tools/openmpi-1.6.5/examples/hello_cxx...Reading symbols from /Users/meredithk/tools/openmpi-1.6.5/examples/hello_cxx.dSYM/Contents/Resources/DWARF/hello_cxx...done.
>>>>> done.
>>>>> 0x00007fff8c1859aa in ?? ()
>>>>> (gdb) bt
>>>>> #0 0x00007fff8c1859aa in ?? ()
>>>>> #1 0x0000000106b73ea0 in ?? ()
>>>>> #2 0x706d6e65706f2f2f in ?? ()
>>>>> #3 0x0000000000000001 in ?? ()
>>>>> #4 0x0000000000000000 in ?? ()
>>>>>
>>>>> This output from gdb was not terribly helpful to me.
>>>>>
>>>>> Karl
>>>>>
>>>>>
>>>>> On Nov 25, 2013, at 8:30 AM, Hammond, Simon David (-EXP) <sdhammo_at_[hidden]<mailto:sdhammo_at_[hidden]>> wrote:
>>>>>
>>>>> We have occasionally had a problem like this when we set LD_LIBRARY_PATH only. On OSX you may need to set DYLD_LIBRARY_PATH instead ( set it to the same lib directory )
>>>>>
>>>>> Can you try that and see if it resolves the problem?
>>>>>
>>>>>
>>>>>
>>>>> Si Hammond
>>>>> Sandia National Laboratories
>>>>> Remote Connection
>>>>>
>>>>>
>>>>> -----Original Message-----
>>>>> From: Meredith, Karl [karl.meredith_at_[hidden]<mailto:karl.meredith_at_[hidden]>]
>>>>> Sent: Monday, November 25, 2013 06:25 AM Mountain Standard Time
>>>>> To: Open MPI Users
>>>>> Subject: [EXTERNAL] Re: [OMPI users] open-mpi on Mac OS 10.9 (Mavericks)
>>>>>
>>>>>
>>>>> I do have these two environment variables set:
>>>>>
>>>>> LD_LIBRARY_PATH=/Users/meredithk/tools/openmpi/lib
>>>>> PATH=/Users/meredithk/tools/openmpi/bin
>>>>>
>>>>> Running mpirun seems to work fine with a simple command, like hostname:
>>>>>
>>>>> $ )mpirun -n 2 hostname
>>>>> meredithk-mac.corp.fmglobal.com<http://meredithk-mac.corp.fmglobal.com>
>>>>> meredithk-mac.corp.fmglobal.com<http://meredithk-mac.corp.fmglobal.com>
>>>>>
>>>>> I am trying to run the simple hello_cxx example from the openmpi distribution, compiled as such:
>>>>> mpic++ -g hello_cxx.cc -o hello_cxx
>>>>>
>>>>> It compiles fine, without warning or error. However, when I go to run the example, it stalls on the MPI::Init() command:
>>>>> mpirun -np 1 hello_cxx
>>>>> It never errors out or crashes. It simply hangs.
>>>>>
>>>>> I am using the same mpic++ and mpirun version:
>>>>> $ )which mpirun
>>>>> /Users/meredithk/tools/openmpi/bin/mpirun
>>>>>
>>>>> $ )which mpic++
>>>>> /Users/meredithk/tools/openmpi/bin/mpic++
>>>>>
>>>>> Not quite sure what else to check.
>>>>>
>>>>> Karl
>>>>>
>>>>>
>>>>> On Nov 23, 2013, at 5:29 PM, Ralph Castain <rhc_at_[hidden]<mailto:rhc_at_[hidden]>> wrote:
>>>>>
>>>>>> Strange - I run on Mavericks now without problem. Can you run "mpirun -n 1 hostname"?
>>>>>>
>>>>>> You also might want to check your PATH and LD_LIBRARY_PATH to ensure you have the prefix where you installed OMPI 1.6.5 at the front. Mac distributes a very old version of OMPI with its software and you don't want to pick it up by mistake.
>>>>>>
>>>>>>
>>>>>> On Nov 22, 2013, at 1:45 PM, Meredith, Karl <karl.meredith_at_[hidden]<mailto:karl.meredith_at_[hidden]>> wrote:
>>>>>>
>>>>>>> I recently upgraded my 2013 Macbook Pro (Retina display) from 10.8 to 10.9. I downloaded and installed openmpi-1.6.5 and compiled it with gcc 4.8 (gcc installed from macports).
>>>>>>> openmpi compiled and installed without error.
>>>>>>>
>>>>>>> However, when I try to run any of the example test cases, the code gets stuck inside the first MPI::Init() call and never returns.
>>>>>>>
>>>>>>> Any thoughts on what might be going wrong?
>>>>>>>
>>>>>>> The same install on OS 10.8 works fine and the example test cases run without error.
>>>>>>>
>>>>>>> Karl
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> users mailing list
>>>>>>> users_at_[hidden]<mailto:users_at_[hidden]>
>>>>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>>>>
>>>>>> _______________________________________________
>>>>>> users mailing list
>>>>>> users_at_[hidden]<mailto:users_at_[hidden]>
>>>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>>>
>>>>> _______________________________________________
>>>>> users mailing list
>>>>> users_at_[hidden]<mailto:users_at_[hidden]>
>>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>>>
>>>>> _______________________________________________
>>>>> users mailing list
>>>>> users_at_[hidden]<mailto:users_at_[hidden]>
>>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>>>
>>>>> _______________________________________________
>>>>> users mailing list
>>>>> users_at_[hidden]
>>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>>
>>>> _______________________________________________
>>>> users mailing list
>>>> users_at_[hidden]
>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>
>>> _______________________________________________
>>> users mailing list
>>> users_at_[hidden]
>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>
>> _______________________________________________
>> users mailing list
>> users_at_[hidden]
>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>
>
> --
> Jeff Squyres
> jsquyres_at_[hidden]
> For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/
>
> _______________________________________________
> users mailing list
> users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/users