[maker-devel] Maker-Error when started with IMPI : CORRECTED MAIL : SEE THIS ONE

[Carson Holt]

I was able to replicate the error as so —>

1. Intel MPI installed on CentOS kernel 6 (MPI works fine)
2. Upgrade to kernel 7 without reinstalling and Intel MPI reports the same error as reported by the user.
3. After recompiling Intel MPI on kernel 7 the error goes away.

The proof that there is an issue with your Intel MPI installation is in this command —>
/opt/bwhpc/common/compiler/intel/compxe.2016.4.258/impi/5.1.3.223/intel64/bin/mpiexec -n 2 echo Hello

That command is simply trying to get mpiexec to launch “echo Hello” internally. And it failed. It’s as simple as that.

Thanks,
Carson




> On Mar 6, 2017, at 1:21 AM, Rainer Rutka <rainer.rutka at uni-konstanz.de> wrote:
> 
> 
> Hi Carson.
> 
> Again thank you for your response.
> 
> But - sorry to say - it's not possible our MPI is corrupt.
> We have approx. 1.500 users working on our bwUniCluster so far. 95 %
> of these users use MPI. And: All our other software (see:
> 
> cis-hpc.uni-konstanz.de )
> 
> is running with our implementations of IMPI/OMPI without any
> issues.
> 
> :-()
> 
> 
> Am 02.03.2017 um 18:41 schrieb Carson Holt:
>> This command -> /opt/bwhpc/common/compiler/intel/compxe.2016.4.258/impi/5.1.3.223/intel64/bin/mpiexec -n 2 echo Hello
>> 
>> All that command does is start the launcher and print “Hello”. So since it failed, it means the issue is with your MPI installation (i.e. Intel MPI itself). It would have to be reinstalled and recompiled. I would not be surprised if the issues with the other MPI flavors you tried were for the same reason.  They were installed for one architecture/compiler/library set, but you are running them on another one. So they always fail.
>> 
>> The second command was an alternate launcher, but it relys on the same underlying libraries as the first one. So if the first one failed, the second one may fail (it may just happen later on).
>> 
>> 
>> So the issue boils down to one thing —> Your MPI is the issue. You need to reinstall/reconfigure and once you can get your MPI working, you can move onto trying MAKER.
>> 
>> Thanks,
>> Carson
>> 
>> 
>> 
>>> On Mar 2, 2017, at 1:41 AM, Rainer Rutka <rainer.rutka at uni-konstanz.de> wrote:
>>> 
>>> Hi Carson!
>>> 
>>> Am 02.03.2017 um 01:43 schrieb Carson Holt:
>>>> Try this command —> /opt/bwhpc/common/compiler/intel/compxe.2016.4.258/impi/5.1.3.223/intel64/bin/mpiexec -n 2 echo Hello
>>>> Then this command —> /opt/bwhpc/common/compiler/intel/compxe.2016.4.258/impi/5.1.3.223/intel64/bin/mpiexec -n 2 /opt/bwhpc/common/bio/maker/2.31.8_impi/bin/maker -h
>>> Same error(s).
>>> 
>>>> If both of these fail, there is the chance that the Intel MPI you are using was compiled on a different architecture than the one you are launching it on. In that case the failure indicates a need to reinstall Intel MPI for that architecture.
>>> Yes, they fail.
>>> 
>>>> The following may or may not work if the first two fail:
>>>> Then this command —> /opt/bwhpc/common/compiler/intel/compxe.2016.4.258/impi/5.1.3.223/intel64/bin/mpiexec.hydra -n 2 echo Hello
>>> WORKS FINE!
>>> 
>>>> Then this command —> /opt/bwhpc/common/compiler/intel/compxe.2016.4.258/impi/5.1.3.223/intel64/bin/mpiexec.hydra -n 2 /opt/bwhpc/common/bio/maker/2.31.8_impi/bin/maker -h
>>> WORKS!
>>> 
>>>> Also send me this file —> perl/lib/MAKER/ConfigData.pm
>>> Attached to this mail.
>>> 
>>>> Thanks,
>>>> Carson
>>> 
>>> --
>>> Rainer Rutka
>>> University of Konstanz
>>> Communication, Information, Media Centre (KIM)
>>> * High-Performance-Computing (HPC)
>>> * KIM-Support and -Base-Services
>>> Room: V511
>>> 78457 Konstanz, Germany
>>> +49 7531 88-5413
>>> <ConfigData.pm>
>> 
> 
> -- 
> Rainer Rutka
> Universität Konstanz
> Kommunikations-, Informations-, Medienzentrum (KIM)
> * KIM Ausbildung
> * Wissenschaftliches Rechnen/bwHPC-C5
> * KIM Basisdienste, KIM Support
> Raum: V511
> 78457 Konstanz
> +49 7531 88-5413
>