[Wien] Wien2k jobs on non-Unix filesystems

[Jeff Spirko]

Wien2k developers,

Recently it has become more common for us to store and/or transport 
Wien2k jobs on non-Unix filesystems (**If you want to know why, see 
below).  These are either networked drives using CIFS/SMB (Windows 
shares) or FAT32 or NTFS partitions.  Wien2k, unfortunately, uses a few 
legitimate Unix filenames which cause problems for CIFS.  While it is 
possible to zip or tar.gz the job before transport/storage, this is 
inconvenient, especially for large jobs.

The problems encountered are:  (1) The :log has a colon in it, making it 
illegal in CIFS.  (2) The case.inm and case.inM use the same letters 
with upper and lower case, making them the same file in CIFS (mixed case 
is allowed, but not distinguished).  This affects some other mixer/mini 
files.

Is there any chance of having a great renaming in the Wien2k scripts, 
such that :log is named case.log (and :parallel, etc.), and the mini 
files duplicate the M (i.e. case.inMM for the mini input)?  If this is 
done, I'd suggest renaming everything case.* (even .machines and 
.time1_1, etc.), but this would only be for convenience, not 
functionality.  Making these small changes would allow even running 
Wien2k jobs or analysis scripts on CIFS filesystems, if necessary.





** Why would one store Unix files on a non-Unix filesystem?  We may be 
getting a 7 TB RAID array that will be NFS mounted to the server (no 
problems) but CIFS (i.e. Windows share) mounted to our desktop systems 
to allow user authentication (problems, see below).  Another case is 
using large USB hard drives, formatted for Windows.  It may be 
inconvenient to partition for Linux, guessing what sizes will be needed 
and wasting some space.

-- 
Jeff Spirko   spirko at lehigh.edu   spirko at gmail.com   WD3V   |=>

The study of non-linear physics is like the study of non-elephant biology.