Re: [zigo@debian.org: Availability of Debian official cloud images from debian.org]
Le Fri, Mar 15, 2013 at 09:25:55AM +0100, olivier.sallou@codeless.fr a écrit :
>
> Le 3/15/13 8:13 AM, Andreas Tille a écrit :
> > Hi,
> >
> > any volunteer to join this effort to create such images featuring
> > med-cloud packages?
> What would be a med-cloud image ? (but an image containing med-cloud
> packages)
Hi all,
the idea behind med-cloud is to install as much tools as possible, while
limiting the size of the image. Hence it is limited to tools that do not
depend on graphical environments, but if some command-line tools would pull
hundreds of megabytes of packages just for themselves, I think that they would
not be fit for the med-cloud metapackage.
Other from this, it should be a standard installation. Ideally we should
improve the support for Blends in tasksel, so that making a "med-cloud" image
would be only the matter of one parameter when creating an image.
Alternatively, a user running a standard image and installing med-cloud should
have everything he needs (except data !) to run simple analysis, including the
analysis of large quantities of sequence reads.
Further usage will show us what adjustment we need. For instance, if /tmp is
in the virtual memory, I expect that some workflow will not work when for
instance sorting multi-gigabyte BED files, etc.
In complemnt with this, it would be great to have tutorials explaining how to
fetch data and make it ready for use with the tools that we distribute. One
solution I know is 'getdata', but did not pass beyond the proof-of-principle
stage. Can biomaj be used on the command line ?
Alternatively or in complement, we can also design a standard file hierarchy
and configure our tools to use it, so that one could prepare separately some
data repositories ready for being used from Debian Med images.
Have a nice week-end,
--
Charles Plessy
Debian Med packaging team,
http://www.debian.org/devel/debian-med
Tsurumi, Kanagawa, Japan
Reply to: