[petsc-dev] failure of new parallel configure

[Satish Balay]

On Tue, 3 Mar 2015, Satish Balay wrote:

> Right not most nightly builds are broken.
> 
> ftp://ftp.mcs.anl.gov/pub/petsc/nightlylogs/archive/2015/03/03/next.html
> 
> It might be the new parallel build has non-deterministic behaviror -
> so issues will be harder to reproduce?
> 
> Perhaps this branch should be reverted from next until its stable?
> 
> Satish
> 
> -----------
> 
> A couple ofissues that can be reproduced:
> 
> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> 1:
> $ ./configure --useThreads=0
> ===============================================================================
>              Configuring PETSc to compile on your system                       
> ===============================================================================
> TESTING: configureExternalPackagesDir from config.framework(config/BuildSystem/config/framework.py:832)                                                       > 
> <hang>

BTW: This hangs in such a way that 'Ctrl-C' doesn't work. I'm having to do 'Ctrl-Z' and then 'kill -9 %1'

Satish

> 
> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> 
> 2: python24 test:
> 
> balay at es^/scratch/balay/petsc(next=) $ ./config/examples/arch-linux-pkgs-opt.py                                                                            
> ===============================================================================
>              Configuring PETSc to compile on your system                       
> ===============================================================================
> TESTING: configureExternalPackagesDir from config.framework(/scratch/balay/petsc/config/BuildSystem/config/framework.py:832)
> TESTING: configureMkdir from config.programs(/scratch/balay/petsc/config/BuildSystem/config/programs.py:22)
> TESTING: configureAutoreconf from config.programs(/scratch/balay/petsc/config/BuildSystem/config/programs.py:44)
> TESTING: configurePrograms from config.programs(/scratch/balay/petsc/config/BuildSystem/config/programs.py:66)
> *******************************************************************************
>         CONFIGURATION CRASH  (Please send configure.log to petsc-maint at mcs.anl.gov)
> *******************************************************************************
> 
> Exception in thread Thread-2:
> Traceback (most recent call last):
>   File "/homes/petsc/soft/linux-Ubuntu_12.04-x86_64/Python-2.4.6/lib/python2.4/threading.py", line 442, in __bootstrap
>     self.run()
>   File "/homes/petsc/soft/linux-Ubuntu_12.04-x86_64/Python-2.4.6/lib/python2.4/threading.py", line 422, in run
>     self.__target(*self.__args, **self.__kwargs)
>   File "/scratch/balay/petsc/config/BuildSystem/config/framework.py", line 1012, in processChildren
>     q.task_done()
> AttributeError: Queue instance has no attribute 'task_done'
> 
> balay at es^/scratch/balay/petsc(next=) $ 
> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> 
> 
> 
> 
> On Mon, 2 Mar 2015, Matthew Knepley wrote:
> 
> > On Mon, Mar 2, 2015 at 7:45 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> > 
> > >
> > >   Oh fuck, never mind, if I get another crash I'll report it.
> > 
> > 
> > Note that I am well aware that this is broken for > 1 thread. This is
> > because we
> > share the compiler/linker objects through the framework. I need to go
> > through
> > and fix that, which will get done soon.
> > 
> > There are other potential breakages in the updating of framework data
> > structures,
> > but these are much less likely to cause problems, and are easily fixable. I
> > have
> > not seen these break when using a bunch of threads, unlike the above.
> > 
> >    Matt
> > 
> > 
> > >
> > >   Barry
> > >
> > > > On Mar 2, 2015, at 6:50 PM, Matthew Knepley <knepley at gmail.com> wrote:
> > > >
> > > > On Mon, Mar 2, 2015 at 6:24 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> > > >
> > > > > On Mar 2, 2015, at 6:06 PM, Matthew Knepley <knepley at gmail.com> wrote:
> > > > >
> > > > > Is this the latest? If not, can you run it after pulling?
> > > >
> > > > git reports already up to date
> > > >
> > > > This log looks like a success. Did you send the wrong one?
> > > >
> > > >   Matt
> > > >
> > > > >
> > > > >   Thanks,
> > > > >
> > > > >      Matt
> > > > >
> > > > > On Mon, Mar 2, 2015 at 6:05 PM, Barry Smith <bsmith at mcs.anl.gov>
> > > wrote:
> > > > >
> > > > > $ ./configure --download-mpich --with-threadsafety --with-openmp
> > > --with-log=0
> > > > >
> > > ===============================================================================
> > > > >              Configuring PETSc to compile on your system
> > > > >
> > > ===============================================================================
> > > > >
> > > ===============================================================================
> > >
> > >                            Trying to download
> > > http://www.mpich.org/static/downloads/3.1.3/mpich-3.1.3.tar.gz for MPICH
> > >
> > >
> > >  ===============================================================================
> > >
> > >
> > >  ===============================================================================
> > >
> > >                            Trying to download
> > > http://ftp.mcs.anl.gov/pub/petsc/externalpackages/mpich-3.1.3.tar.gz for
> > > MPICH
> > >
> > >  ===============================================================================
> > >
> > >
> > >  ===============================================================================
> > >
> > >                            Running configure on MPICH; this may take
> > > several minutes
> > >
> > >  ===============================================================================
> > >
> > >
> > >  ===============================================================================
> > >
> > >                            Running make on MPICH; this may take several
> > > minutes
> > >
> > > ===============================================================================
> > >
> > >
> > >  ===============================================================================
> > >
> > >                            Running make install on MPICH; this may take
> > > several minutes
> > >
> > > ===============================================================================
> > >
> > >                      TESTING: output from
> > > config.setCompilers(config/BuildSystem/config/setCompilers.py:1451)
> > >
> > >
> > > *******************************************************************************
> > > > >          UNABLE to CONFIGURE with GIVEN OPTIONS    (see configure.log
> > > for details):
> > > > >
> > > -------------------------------------------------------------------------------
> > > > >
> > > > >
> > > *******************************************************************************
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > What most experimenters take for granted before they begin their
> > > experiments is infinitely more interesting than any results to which their
> > > experiments lead.
> > > > > -- Norbert Wiener
> > > >
> > > >
> > > >
> > > >
> > > > --
> > > > What most experimenters take for granted before they begin their
> > > experiments is infinitely more interesting than any results to which their
> > > experiments lead.
> > > > -- Norbert Wiener
> > >
> > >
> > 
> > 
> > 
> 
>