Re: [AMBER] antechamber error

From: David Watson <dewatson.olemiss.edu>
Date: Thu, 10 Dec 2009 17:26:00 -0600

On Dec 10, 2009, at 4:34 PM, Mike Hanby wrote:

> Has anyone come up with a workaround for this while we wait for Amber 11? I distribute all software to our workstations via RPMs and have come up with a spec file that I _thought_ resulted in a working Amber and AmberTools distribution.
>
> Apparently, none of the users over the past year have used antechamber (and I'm assuming there are other tools that suffer from the 'hard coding' of AMBERHOME). Sander, pmemd and other executables all work.
>
> At any rate, here's the error reported by one of my users:
>
> $ antechamber -i sustiva.pdb -fi pdb -o sustiva.mol2 -fo mol2 -c bcc -s 2
> Cannot open connect_file /home/builder/rpm/BUILD/amber10/dat/antechamber/CONNECT.TPL in connect(), exit
>
> Obviously, that's the path used during the build, but the software actually gets installed to /opt/amber10.
>
> I'm wondering if anyone has a patch or suggestion to resolve this in the interim? I tried changing DATDIR in configure_at and configure_amber, however that caused build failures because the build attempts to copy files to the location, and only root has permissions to do that.

I would suggest that you export the amber directory via NFS to all of the workstation machines.
If you export /opt/amber10 for instance, have the other machines automount the directory as /opt/amber10.
Your variables won't change because you won't ever have to package up an rpm.
This way you build once and everyone has immediate access to the product.



_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Dec 10 2009 - 15:30:02 PST
Custom Search