[CPMD-list] Upon compilation of cpmd2cube...
Axel Kohlmeyer
axel.kohlmeyer at theochem.ruhr-uni-bochum.de
Thu May 27 10:46:18 CEST 2004
On Thu, 27 May 2004 kohsj at ihpc.a-star.edu.sg wrote:
AK> Thanks Dr. Axel. I should've mentioned that the previous versions of "Makefile" that you attached did not produce the ".nfsXXXXX" files, only the latest one did.
AK>
AK> Any clue as to why this "Makefile" gave those files?
adrian,
this has _nothing_ to do with the makefile itself.
most likely you have accessed (and overwriteen) the file
with two editor sessions simultaneously.
axel.
AK>
AK> Cheers,
AK> Adrian.
AK>
AK>
AK>
AK> ----- Original Message -----
AK> From: Axel Kohlmeyer <axel.kohlmeyer at theochem.ruhr-uni-bochum.de>
AK> Date: Thursday, May 27, 2004 4:22 pm
AK> Subject: Re: [CPMD-list] Upon compilation of cpmd2cube...
AK>
AK>
AK> On Thu, 27 May 2004 kohsj at ihpc.a-star.edu.sg wrote:
AK>
AK> AK> Dear Dr. Axel,
AK> AK>
AK>
AK> AK> Is there any way that one can remove the .nfsXXXXX files, and
AK> AK> subsequently, the "cpmd2cube-v0.1.1" folder?
AK>
AK> the explanation should have made it clear, that you simply
AK> have to terminate _all_ processes accessing files in that directory,
AK> including editors, file viewers, file managers, directory browsers etc.
AK>
AK> axel.
AK>
AK> AK>
AK> AK> Thanks again!
AK> AK>
AK> AK>
AK> AK>
AK> AK> AK> My IT man suggested that it could probably be:
AK> AK> AK>
AK> AK> AK>
AK> AK> AK> "Since you are the one who create the folder, you shall
AK> have the
AK> AK> right to
AK> AK> AK> remove anything in the folder. However, you claim that you
AK> AK> delete the
AK> AK> AK> file and it come back again. This is mostly likely due to
AK> your
AK> AK> program.AK> You may have write a program and this program may
AK> need
AK> AK> to write
AK> AK> AK> something in this folder everytime it is running and
AK> therefore
AK> AK> the file
AK> AK> AK> come back after you delete it. I just make a guess and I do
AK> not
AK> AK> have any
AK> AK> AK> evidence to support my point. Hope that clear your doubt."
AK> AK>
AK> AK> well, this is close to the explanation. the reason for the
AK> .nfsXXXXX
AK> AK> filesis 'feature' of NFS in order to emulate posix style file
AK> access
AK> AK> semantics.to explain: if you delete a file on a posix/unix
AK> machine,
AK> AK> the file is
AK> AK> initially only removed from the directory listing but not
AK> deleted if
AK> AK> thereis still a process, that has an open file descriptor
AK> pointing
AK> AK> to that
AK> AK> file. only if the last process accessing the file has ended,
AK> that
AK> AK> file is
AK> AK> really deleted. this is btw a very convenient way of creating
AK> AK> 'invisible'scratch files, that get automatically deleted, when
AK> the
AK> AK> process terminates
AK> AK> or dies (you open(2)/fopen(3) a file with a unique name for
AK> writing
AK> AK> and
AK> AK> immediately unlink(2) it). back to the .nfsXXXX file: the NFS
AK> AK> filesystem,however is basically 'stateless', i.e. the
AK> connection to
AK> AK> the NFS server
AK> AK> can be interrupted any time and later reconnected (in some
AK> cases it
AK> AK> can
AK> AK> even survive an intermediate reboot with replacing the hardware
AK> and
AK> AK> resizing the filesystem). so to implement the 'file-stays-even-
AK> if-
AK> AK> deleted'semantics, the nfs server daemon acts as a proxy for
AK> the
AK> AK> processes
AK> AK> accessing the file, but to make this work even in case of a
AK> reboot, the
AK> AK> .nfsXXXX files are created. the XXXXX is an hash, that allows
AK> the
AK> AK> serverto detect the process needing to access the 'file with no
AK> name'.AK> most frequently this happens if you open a file in an
AK> editor, that
AK> AK> keeps an open file descriptor to the file it opens (or to the
AK> AK> associated backup/undo file).
AK> AK>
AK> AK>
AK> AK>
AK> AK> axel.
AK> AK>
AK>
AK> --
AK>
AK>
AK> =======================================================================
AK> Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at rub.de
AK> Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673
AK> Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045
AK> D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
AK> =======================================================================
AK>
AK>
AK>
AK>
--
=======================================================================
Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at rub.de
Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673
Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045
D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
=======================================================================
More information about the CPMD-list
mailing list