[Container-tools] Tool request from LinuxCon: layer re-use

[Muayyad AlSadi]

In Opensooq.com, it does not matter what is your base image, because we use
ansible instead of Dockerfiles

it would bring any given base image into the desired state (wither it's a
fresh bare base image, or latest milestone release of application)

http://engineering.opensooq.com/manage-your-docker-image-layers-with-ansible/


On Thu, Sep 1, 2016 at 2:15 PM, Daniel J Walsh <dwalsh at redhat.com> wrote:

>
>
> On 08/31/2016 08:10 PM, Josh Berkus wrote:
> > CT team:
> >
> > I was talking to some enterprise container builders at LinuxCon, and one
> > thing which came up as a tooling request was some way to
> > encourage/enforce making developers re-use shared container layers, and
> > track if they had.
> >
> > The basic idea is that it's a "best practice" to make reusable layered
> > containers with foundational technology and re-use them -- among other
> > things, to simplify security updates.  So for example, this would be
> > good inherited container heirarchy for a GeoDjango container:
> >
> >
> > Fedora24
> >    |
> > Python27
> >    |
> > Django
> >    |
> > GeoDjango
> >    |
> > MyGeoDjangoApplication
> >
> > And this would be a poor one:
> >
> > Fedora24
> >    |
> > MyGeoDjangoApplication
> >
> >
> > But AFAIK, there aren't any tools to encourage or check for the first
> > over the second.
> >
> Yes I agree this would be good, but not sure how you would do this
> automatically.  The problem is every app could want to use a different
> set of middle layers.   Currently developers pick which layer they want
> to base on, but they might not even know that the GeoDjango layer
> exists.  Tools like OpenShift could build some business sense into
> this.  Or at least give developers a list of images to select from and a
> description of the content in each image.
>
> atomic diff GeoDjango MyGeoDjangoApplication
>
> Theoretically would give you only the RPMs/packages that are different
> between the two builds so if you did the second, you might be able to
> determine you could have done the first.
>
> _______________________________________________
> Container-tools mailing list
> Container-tools at redhat.com
> https://www.redhat.com/mailman/listinfo/container-tools
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/container-tools/attachments/20160901/b5196832/attachment.htm>