Those got moved to GitHub from readthedocs.io, meaning links pointing to specific pages return a 404, so this fixes that along with skipping the redirect for the links that just point to the main page.
4.0 KiB
nav_order |
---|
4 |
Compose server
{: .no_toc }
- TOC {:toc}
Using higher level build tooling
Originally rpm-ostree compose tree
was intended to be a "high level" tool,
but that didn't work out very well in practice. Today, you should consider
it as a low level tool. For example, most people that want to generate
OSTree commits also want to generate bootable disk images, and rpm-ostree
has nothing to do with that.
One example higher level tool that takes care of both OSTree commit generation and bootable disk images is coreos-assembler.
Background on managing an OSTree repository
Before you get started, it's recommended to read (at least) these two sections of the OSTree manual:
Generating OSTree commits from a CentOS base
First, you'll need a copy of rpm-ostree
. The current recommendation is to use
a privileged container, but you can also install rpm-ostree
directly to a
physical or virtual machine.
It's included in the package collection for Fedora, and there are CentOS Core packages as well as bleeding edge CentOS builds.
You can create a privileged container with e.g. podman
via: podman run --privileged registry.fedoraproject.org/fedora:27 ...
. However you create the
environment, run yum -y install rpm-ostree
.
A good first thing to try would be using the CentOS Atomic Host metadata to generate a custom host.
One time setup, where we clone the git repository, then make two OSTree repos, one for doing builds, one for export via HTTP:
# mkdir /srv/centos-atomic
# cd /srv/centos-atomic
# git clone https://github.com/CentOS/sig-atomic-buildscripts -b downstream
# mkdir build-repo
# ostree --repo=build-repo init --mode=bare-user
# mkdir repo
# ostree --repo=repo init --mode=archive
We'll also want to cache downloaded RPMs:
# mkdir cache
Running rpm-ostree compose tree
This program takes as input a manifest file that describes the target system, and commits the result to an OSTree repository.
The input format is a JSON "treefile". See examples in
api-doc/treefile-examples
. More real-world examples include the manifest
for
Fedora Atomic and
CentOS Atomic.
If you're doing this multiple times, it's strongly recommended to create a cache directory:
# rpm-ostree compose tree --unified-core --cachedir=cache --repo=/srv/centos-atomic/build-repo sig-atomic-buildscripts/centos-atomic-host.json
This will download RPMs from the referenced repos, and commit the result to the
OSTree repository, using the ref named by ref
.
Once we have that commit, let's export it:
# ostree --repo=repo pull-local build-repo centos-atomic-host/7/x86_64/standard
You can tell client systems to rebase to it by combining ostree remote add
,
and rpm-ostree rebase
on the client side.