Re: [AMBER] ambertools14 fresh installation fails

From: Jason Swails <jason.swails.gmail.com>
Date: Tue, 27 Jan 2015 17:12:25 -0500

On Tue, Jan 27, 2015 at 3:35 PM, Fabian Gmail <fabian.glaser.gmail.com>
wrote:

> Thanks Dac Jason,
>
> I did removed completely the amber14 and amber12 old dirs before
> attempting a fresh download and installation and I do have a
> $AMBERHOME/.patches dir... Which I did not touch.
>

​Untarring a fresh AmberTools14.tar.bz2 or AmberTools13.tar.bz2 will not
create a .patches directory. It is created the first time update_amber
runs. A simple thing to do would be remove the .patches directory, but its
very existence indicates that something is weird with your install process,
and I don't know what.

​Where did you get the tarball? Downloaded from the Amber website? Does
the md5sum match what the Amber website says it should?
(589cc9bd17b15ce3f9604b0bdeea3bab)
 And you're sure you removed "amber14" and "amber12" before trying this?
Is it possible you ran a command you thought removed everything that really
didn't? (The only example here that comes to mind is running " rm -fr
amber14/* " instead of " rm -fr amber14 "). By default, I think "rm -fr
amber14/*" will not match any "hidden" dot-directories inside amber14 (like
.patches).

I've reached the point of pure conjecture at this point, since there must
be *some* detail of your process that you are not reporting here (perhaps
you don't even know that what you are doing is "different") that is causing
these issues. The code in question (i.e., the part of update_amber that
checks for existing updates) is quite robust and platform-independent (and
routinely tested/used on Mac OS X, anyway), and the contents of the Amber
release tarball are completely platform-independent. This almost
completely precludes the root cause of this problem being Amber


> ​​
> Should I delete it too or its content (the .patches ) before and trying
> again???
>

​Delete everything (the whole existing $AMBERHOME directories) and try
extracting a fresh AmberTools14.tar.bz2 file whose md5sum you have verified
matches the one I printed above.

HTH,
Jason

-- 
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Jan 27 2015 - 14:30:02 PST
Custom Search