Open MPI logo

Open MPI User's Mailing List Archives

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

Subject: Re: [OMPI users] mpirun with WMI on Windows 7
From: Shiqing Fan (fan_at_[hidden])
Date: 2010-02-04 06:35:39


r22549 should solve the mpirun crash problem, it's just few lines
change, you could try to modify the source file, or I can send you a
patch. But the mpirun might still hang due to a tcp connection failure,
I made a ticket for this problem:
https://svn.open-mpi.org/trac/ompi/ticket/2231 .

Thanks,
Shiqing

Charles Shuller wrote:
> I get the same behavior regardless of which machine I execute mpirun from.
>
> The security settings I altered were to add the user to both DCOM
> Access and Launch with all options set to allow.
> In WMI, I add the user to the CIMV2 namespace and grant all permissions
> except edit security.
> Windows firewall is turned off on both machines.
>
> Did I leave anything out?
>
> Thanks!
>
>
> Charles
>
> On 2/2/2010 12:49 PM, Shiqing Fan wrote:
>
>> Hi Charles,
>>
>> I've downloaded a fresh Open MPI 1.4.1 tarball, and built a clean
>> installer. After installing it onto two machines, everything worked
>> for me, I tested to launch notedpad and also other simple MPI
>> applications like barrier, they all worked. So there still might be
>> some settings got wrong on your environment.
>>
>> Could you try to run the application on both machine locally? e.g.
>> "mpirun -np 2 app.exe", probably with full path to the application,
>> just to make sure that the application could be started locally on
>> each machine. And try also "mpirun -np 2 -host host1,host2 app.exe"
>> from both machine, i.e. from both direction.
>>
>>
>> Regards,
>> Shiqing
>>
>>
>> Charles Shuller wrote:
>>
>>> The connection between nodes should be TCP/IP. I am currently using
>>> Open MPI 1.4.1.
>>>
>>> I've attached the output of ompi_info as a text file.
>>>
>>>
>>> Thanks!
>>>
>>>
>>> Charles
>>> On 2/2/2010 11:46 AM, Shiqing Fan wrote:
>>>
>>>
>>>> Hi Charles,
>>>>
>>>> It seems not a WMI problem, because the remote orted has already been
>>>> launched, and only that part was done by WMI.
>>>>
>>>> What connection do you have between the nodes, tcp? Could you provide
>>>> the version information of Open MPI, or just the output of ompi_info?
>>>> so that I can take a close look.
>>>>
>>>>
>>>> Thanks,
>>>> Shiqing
>>>>
>>>>
>>>> Charles Shuller wrote:
>>>>
>>>>
>>>>> No messages on the command prompt.
>>>>>
>>>>> When I executed mpirun to launch notepad on the remote machine, It
>>>>> crashes again.
>>>>>
>>>>> No information is ever printed to the command line unless I enter a
>>>>> bad
>>>>> password.
>>>>>
>>>>> The very first time I attempt to use mpirun to launch a process on the
>>>>> remote machine, I get an indefinite hang (I let it run for several
>>>>> hours
>>>>> yesterday). Subsequently I get an abend dialog in about 3 seconds.
>>>>>
>>>>> My MPI application (Which just calls init and finalize) is at C:\bin
>>>>> which is in the system path on both machines, this is also the bin
>>>>> directory for the openmpi package.
>>>>>
>>>>> Is there anyway I can turn on logging, or do I need to go through and
>>>>> insert debug statements myself and recompile?
>>>>>
>>>>>
>>>>>
>>>>> Thanks!
>>>>>
>>>>>
>>>>> Charles
>>>>>
>>>>> On 2/2/2010 11:17 AM, Shiqing Fan wrote:
>>>>>
>>>>>
>>>>>
>>>>>> Hi Charles,
>>>>>>
>>>>>> On the local machine, which can be considered also as the "head
>>>>>> node",
>>>>>> no orted will be launched, the mpirun itself works as it locally.
>>>>>>
>>>>>> Did you see any error message on the command prompt? That would be
>>>>>> very helpful.
>>>>>>
>>>>>> To do a simple test, just try to launch notepad on remote node:
>>>>>> mpirun
>>>>>> -np 1 -host host1 notepad.exe , and this will do the same thing as
>>>>>> you run the wmic command line.
>>>>>>
>>>>>> If that works, it might mean that, you didn't copy your MPI
>>>>>> application onto the remote node, which should present in the same
>>>>>> path as on all working nodes, for example, on working nodes, your
>>>>>> application could be placed at: D:\tests\app\app.exe (You have to do
>>>>>> so, because the wmi impersonate level doesn't support network share
>>>>>> yet. I'm still hacking on this to break the limit.). And then you can
>>>>>> run the mpirun command line supplied with the application full
>>>>>> path or
>>>>>> directly run it under the application path.
>>>>>>
>>>>>>
>>>>>> Regards,
>>>>>> Shiqing
>>>>>>
>>>>>>
>>>>>>
>>>>>> Charles Shuller wrote:
>>>>>>
>>>>>>
>>>>>>> Thanks Shiqing!
>>>>>>>
>>>>>>> Unfortunately, it still doesn't work, but I've got more info.
>>>>>>>
>>>>>>> I can use wmic to start an application on the remote machine, but
>>>>>>> that application does not start in the current login process
>>>>>>> (notepad.exe starts, but I have to ask task manager to show all
>>>>>>> processes to find it, even though I'm currently logged in as the
>>>>>>> same
>>>>>>> user). I believe this is expected behavior, please let me know if
>>>>>>> it's not.
>>>>>>>
>>>>>>> When using mpirun, I can verify that orted starts on the remote
>>>>>>> machine, but the crash or hang appears to happen before the
>>>>>>> application starts execution. Oddly, orted does not appear to
>>>>>>> start
>>>>>>> on the local machine. Logs all refer to mpirun crashing.
>>>>>>>
>>>>>>>
>>>>>>> Cheers!
>>>>>>>
>>>>>>> Charles
>>>>>>>
>>>>>>> On 1/29/2010 2:56 AM, Shiqing Fan wrote:
>>>>>>>
>>>>>>>
>>>>>>>> Hi Charles,
>>>>>>>>
>>>>>>>> You don't need to install anything, but just a few security setting
>>>>>>>> has to be correctly configured. Here are two links might be helpful
>>>>>>>> (will be added into README.WINDOWS too):
>>>>>>>> http://msdn.microsoft.com/en-us/library/aa393266(VS.85).aspx
>>>>>>>> http://community.spiceworks.com/topic/578
>>>>>>>>
>>>>>>>> On the other hand, in order to check if WMI is working between the
>>>>>>>> nodes, you can try with command:
>>>>>>>>
>>>>>>>> C:\>wmic /node:192.168.0.1 /user:username process call create
>>>>>>>> notepad.exe
>>>>>>>>
>>>>>>>> the ip has to be the remote computer ip address, and the user name
>>>>>>>> is which you use on remote computer. This command line will simply
>>>>>>>> launch a non-interactive notepad (no GUI) on remote node using WMI,
>>>>>>>> if it is successful, you should be able to see a notepad process in
>>>>>>>> Task Manager or Process Viewer, and that also means mpirun will
>>>>>>>> work
>>>>>>>> through WMI.
>>>>>>>>
>>>>>>>> Could you check with the above command, and possibly tell me the
>>>>>>>> return value, so that I can help you to make it work.
>>>>>>>>
>>>>>>>>
>>>>>>>> Regards,
>>>>>>>> Shiqing
>>>>>>>>
>>>>>>>>
>>>>>>>> Charles Shuller wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>> When attempting to launch an application on both local and remote
>>>>>>>>> windows7 hosts, mpirun either hangs indefinately or abends.
>>>>>>>>>
>>>>>>>>> The application executes correctly on both machines, when only
>>>>>>>>> launched
>>>>>>>>> on a single host.
>>>>>>>>>
>>>>>>>>> I believe mpirun is using WMI, README.WINDOWS indicates that this
>>>>>>>>> is the
>>>>>>>>> case if I don't have the CCP toolkit and SDK installed, which I
>>>>>>>>> don't. Additionally, I have encountered and resolved some
>>>>>>>>> security
>>>>>>>>> issues
>>>>>>>>> following this assumption.
>>>>>>>>>
>>>>>>>>> Any advice is welcome. I'm not married to WMI, so if the
>>>>>>>>> solution is
>>>>>>>>> "install something else" I'm great with that.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Thanks!
>>>>>>>>>
>>>>>>>>> Charles
>>>>>>>>> _______________________________________________
>>>>>>>>> users mailing list
>>>>>>>>> users_at_[hidden]
>>>>>>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>
>>>
>>
>
>

-- 
--------------------------------------------------------------
Shiqing Fan                          http://www.hlrs.de/people/fan
High Performance Computing           Tel.: +49 711 685 87234
  Center Stuttgart (HLRS)            Fax.: +49 711 685 65832
Address:Allmandring 30               email: fan_at_[hidden]    
70569 Stuttgart