[maker-devel] Need your help on maker pipeline

Carson Holt carsonhh at gmail.com
Mon Nov 27 12:45:57 MST 2017


The parameters needed to get OpenMPI to work with MAKER are described in the …/maker/INSTALL file (specifically look at LD_PRELOAD and -mca btl ^openib) —> 

!!IMPORTANT!!
MAKER is not compatible with MVAPICH2. Use OpenMPI or MPICH. If using MPICH, make
sure to enable shared libaries during installation (this is not the default). If
using OpenMPI, make sure to set LD_PRELOAD to the location of libmpi.so before
even trying to install MAKER. It must also be set before running MAKER (or any
program that uses OpenMPI's shared libraries), so it's best just to add it to
your ~/.bash_profile. (i.e. export LD_PRELOAD=/usr/local/openmpi/lib/libmpi.so).


1.  Say yes to the 'configure for MPI' question when running 'perl Build.PL' in
    step 1 of the EASY INSTALL.

2.  Give path to 'mpicc'. Note to make sure you do not give the path to 'mpicc'
    from another MPI flavor that might be installed on your system.

3.  Give path to the folder containing 'mpi,h'. Note to make sure you do not
    give the path to a folder from another MPI flavor that might be installed
    on your system. Mixing MPI flavors for 'mpicc' and 'mpi.h' will cause
    failures. Make sure to read and confirm the auto-detected paths.

4.  Finish installation according to steps 2-4 of the EASY INSTALL

    Note: For OpenMPI you may also want to set OMPI_MCA_mpi_warn_on_fork=0 in
    your ~/.bash_profile to turn off certain nonfatal warnings.

    Note: If jobs hang or freeze when using mpiexec under OpenMPI try adding
    the '-mca btl ^openib' flag to mpiexec command when running MAKER.

        Example: mpiexec -mca btl ^openib -n 20 maker



Then to diable the forks warning, just add the parameter --mca mpi_warn_on_fork 0 to the mpiexec options as described in the warning.

How to run with OpenMPi has also been covered extensively ibn the MAKER list archives and more detail can be found there —> https://groups.google.com/forum/#!searchin/maker-devel/openmpi%7Csort:date <https://groups.google.com/forum/#!searchin/maker-devel/openmpi|sort:date>

Thanks,
Carson



> On Nov 27, 2017, at 6:18 AM, HAI WANG <wanghai01 at caas.cn> wrote:
> 
> Dear Professor Yandell,
>  
> I am Hai Wang, a visiting scholar in Cornell University. I am sorry to bother you, but I really need your help. I am now using the maker pipeline to annotate a maize genome. The installation of maker, openmpi and other software should be OK since I’ve successfully run maker on your example data.
>  
> But when I ran maker on my own maize genome, I always got the following error:
>  
>  
> A process has executed an operation involving a call to the
> "fork()" system call to create a child process.  Open MPI is currently
> operating in a condition that could result in memory corruption or
> other system errors; your job may hang, crash, or produce silent
> data corruption.  The use of fork() (or system() or other calls that
> create child processes) is strongly discouraged.
>  
> The process that invoked fork was:
>  
>   Local host:          [[21269,1],0] (PID 12537)
>  
> If you are *absolutely sure* that your application will successfully
> and correctly survive a call to fork(), you may disable this warning
> by setting the mpi_warn_on_fork MCA parameter to 0.
> --------------------------------------------------------------------------
> --------------------------------------------------------------------------
> mpiexec noticed that process rank 32 with PID 0 on node fat1 exited on signal 11 (Segmentation fault).
> --------------------------------------------------------------------------
>  
> Could you please help me with this issue? Or is there a way that I can resume this job when it stops? Thank you very much!
>  
> Best,
> Hai Wang
>  
> _______________________________________________
> maker-devel mailing list
> maker-devel at box290.bluehost.com <mailto:maker-devel at box290.bluehost.com>
> http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org <http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://yandell-lab.org/pipermail/maker-devel_yandell-lab.org/attachments/20171127/e1e99174/attachment-0003.html>


More information about the maker-devel mailing list