[Bio-Linux] BioLinux8 Dockerfile (Tony Travis)

[Brad Chapman]

Steve and Tony;
I agree with this approach. This is essentially what we are doing with
CloudBioLinux inside bcbio: providing a big ol' container with a lot of
useful tools related to next-gen sequencing:

https://github.com/chapmanb/bcbio-nextgen/blob/master/Dockerfile

This uses both Bio-Linux and other installers like Homebrew cobbled
together with CloudBioLinux.

There is definitely an issue of size as you add more tools but if you
want to run them together it's easier to manage and distribute a single
instance rather than trying to chain multiple images. So somewhere in
between a single tool and all the tools in the world is the right place
to be. It would be cool to see what you can come up with for a Bio-Linux
distribution,
Brad

> Dear Tony,
>
> I'm not wanting to create a full Bio-Linux installation in a container. As
> you say, that would be contrary to the purpose of containers. I'm more
> interested in creating a cut-down Bio-Linux-eqsue Docker image that has all
> the relevant command line tools installed for undertaking bioinformatics
> analyses, and access to the Bio-Linux 8 repos for installing additional
> tools as necessary.
>
> Another way of doing this would be to simply use the ubuntu base image, add
> the Bio-Linux 8 repos, pin any GUI packages in apt with -1 priority, and
> install all command line packages as part of the Dockerfile configuration.
> Just trying to think of the easiest way to approach it.
>
> I've followed Brad's work closely, so be interested to hear his take on
> this too.
>
> Cheers,
>
> Steve
>
> On 16 December 2014 at 12:00, <bio-linux-request at nebclists.nerc.ac.uk>
> wrote:
>>
>> Message: 1
>> Date: Mon, 15 Dec 2014 12:24:01 +0000
>> From: Tony Travis <tony.travis at abdn.ac.uk>
>> To: <bio-linux at nebclists.nerc.ac.uk>
>> Cc: Brad Chapman <chapmanb at 50mail.com>
>> Subject: Re: [Bio-Linux] BioLinux8 Dockerfile
>> Message-ID: <548ED2E1.4050301 at abdn.ac.uk>
>> Content-Type: text/plain; charset="windows-1252"
>>
>> On 14/12/14 14:40, Booth, Timothy G. wrote:
>> > [...]
>> > To answer your question, I'd suggest the following approach - take a
>> > fresh Bio-Linux VM (ie. the .ova download) then remove the Gnome and
>> > X libs packages.  Lots of packages that depend on them will be
>> > removed.  Then you can diff the master package list from the upgrade
>> > script and the packages you have left on the system to get your
>> > blacklist.  I'm sure there are other approaches but this is the one
>> > that springs to mind.
>>
>> Hi, Tim and Steve.
>>
>> I think you're missing the point about Docker if you're trying to create
>> a Bio-Linux container. Docker is a lightweight container that relies on
>> the host OS to provide kernel services and system daemons.
>>
>> Bio-Linux is more appropriate as a host OS for Docker. I don't really
>> see the point of trying to run a full Bio-linux in a Docker container
>> because the overhead will be such that you might as well be using a VM.
>>
>> Docker is most useful for providing an OS independent environment for
>> applications like e.g. GATK that have specific version dependencies that
>> are quite difficult to reconcile with versions of programs that are
>> installed on a particular host. Docker provides portability of an
>> application between host and OS's by encapsulating these dependencies.
>>
>> Having said that, I'm currently trying to use Bio-Linux in a user-level
>> "fakechroot" under 'Scientific' Linux (CentOS 6.3) on our HPC cluster
>> because our out-sourced IT support company think Docker is too serious a
>> security threat to be used on a 'production' server. I've pointed out
>> that both AWS and Rackspace are offering Docker instances on their own
>> cloud services 'securely'. I'm experimenting with Docker on my own PC.
>>
>> [I've CC'ed this to Brad to get his take on using Docker for Bio-Linux]
>>
>> Bye,
>>
>>   Tony.
>>
>
> -- 
>
> Steve Moss
> about.me/gawbul
> [image: Steve Moss on about.me]
>   <http://about.me/gawbul>