i'm putting together some docker material for a client and, as
always, i'm a real stickler for accurate terminology and i just ran
across something that confuses me just a bit.
i already know that docker creates a final, runnable image by
starting with some base image (fedora, ubuntu, whatever), then
adding other layers on top of that to come up with the final runnable
image.
an example in a book i'm looking at -- "the docker book" by james
turnbull -- uses the following example of lower, read-only layers, on
top of which is placed the final read-write layer when the image is
run:
* writeable (initially empty) layer
* apache
* emacs
* ubuntu (base image)
so this clearly represents an image that was created by starting with
a base ubuntu image, then adding an emacs layer, then adding an apache
layer on top of that. but here's my confusion.
up to now, i thought that the definition of an "image" was the
accumulation of the filesystem layers for each layer; that is, in the
picture above, i would have said that there are three images:
* ubuntu base image
* ubuntu base image + emacs
* ubuntu base image + emacs + apache
see what i mean? everything i'd read up to now seemed pretty clear
that an "image" was the combination of some filesystem layer all the
way to the bottom. but turnbull defines an image as simply the
contents of each filesystem layer, so turnbull's list of images would
be:
* ubuntu base image
* emacs filesystem/image
* apache filesystem/image
i just like to be precise ... how do others on this list see this?
thanks.
rday
--
========================================================================
Robert P. J. Day Ottawa, Ontario, CANADA
http://crashcourse.ca
Twitter: http://twitter.com/rpjday
LinkedIn: http://ca.linkedin.com/in/rpjday
========================================================================