[Pvfs2-users] Test problems with hdf5 and pvfs2

[Tobias Brandvik]

On Jan 15, 2008 7:42 PM, Robert Latham <robl at mcs.anl.gov> wrote:
> I'm super-excited to see someone else besides a ROMIO developer use
> the file system prefix!
>
> Do you happen to have PVFS mounted also?

I'm relatively new to parallel filesystems (new to filesystems in
general actually), but I take it this means using PVFS through the VFS
interface by building the kernel modules and so on? If so, no.

> > Proc 0: *** PHDF5 ERROR ***
> >         Assertion (H5Fcreate) failed at line  126 in t_chunk_alloc.c
>
> I strongly suspect that the HDF5 test here assumes that it can do
> 'open(/mnt/pvfs2/filename)' -- i.e. access the HDF5_PARAPREFIX
> directory through the VFS.  I hammer those guys on this point time and
> time again.  It looks like this new test is buggy with respect to
> ROMIO filenames.

So HDF5 itself should be ok - it's only the test itself that's buggy?

> I'll take this up with the HDF5 guys and we'll see if we can get a fix
> to you.  In the meantime, try running all the other tests (run with
> -exclude calloc)  and let us know if all the other tests still pass.

I had to skip a few tests to get everything working:

Testing  -- fapl_mpio duplicate (mpiodup)
Testing  -- fapl_mpiposix duplicate (posixdup)
Testing  -- dataset using split communicators (split)
Testing  -- dataset independent write (idsetw)
Testing  -- dataset independent read (idsetr)
Testing  -- dataset collective write (cdsetw)
Testing  -- dataset collective read (cdsetr)
Testing  -- extendible dataset independent write (eidsetw)
Testing  -- extendible dataset independent read (eidsetr)
Testing  -- extendible dataset collective write (ecdsetw)
Testing  -- extendible dataset collective read (ecdsetr)
Testing  -- extendible dataset independent write #2 (eidsetw2)
Testing  -- chunked dataset with none-selection (selnone)
Skipping -- parallel extend Chunked allocation on serial file (calloc)
Skipping -- parallel read of dataset written serially with filters (fltread)
Skipping -- compressed dataset collective read (cmpdsetr)
Testing  -- multiple datasets write (ndsetw)
Testing  -- multiple groups write (ngrpw)
Testing  -- multiple groups read (ngrpr)
Testing  -- compact dataset test (compact)
Testing  -- collective group and dataset write (cngrpw)
Testing  -- independent group and dataset read (ingrpr)
Skipping -- big dataset test (bigdset)
Testing  -- dataset fill value (fill)
Testing  -- simple collective chunk io (cchunk1)
Testing  -- noncontiguous collective chunk io (cchunk2)
Testing  -- multi-chunk collective chunk io (cchunk3)
Skipping -- collective irregular contiguous write (ccontw)
Skipping -- collective irregular contiguous read (ccontr)
Testing  -- collective irregular simple chunk write (cschunkw)
Testing  -- collective irregular simple chunk read (cschunkr)
Testing  -- collective irregular complex chunk write (ccchunkw)
Testing  -- collective irregular complex chunk read (ccchunkr)
Testing  -- I/O mode confusion test -- hangs quickly on failure (I/Omodeconf)

All tests were successful.

Thanks for your help!

Tobias