| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
| |
We now handle the two pieces of upgrade that require a node evac in the
same play. (docker, and node itself)
|
| |
|
| |
|
|
|
|
|
|
| |
Docker 1.10 is not widely available in RHEL / CentOS yet, lets remove
the restriction for users of master / origin. We can revert as soon as
it's available publically.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adds a separate playbook for Docker 1.10 upgrade that can be run
standalone on a pre-existing 3.2 cluster. The upgrade will take each
node out of rotation, and remove *all* containers and images on it, as
this is reportedly faster and more storage efficient than performing the
in place 1.10 upgrade.
This process is integrated into the 3.1 to 3.2 upgrade process.
Normal config playbooks now become 3.2 only, and require Docker 1.10.
Users of older environments will have to use an appropriate
openshift-ansible version.
Config playbooks no longer are in the business of upgrading or
downgrading docker.
|
|\
| |
| | |
Don't pull cli image when we're not containerized
|
| | |
|
|\ \
| | |
| | | |
Fixes for openshift_docker_hosted_registry_insecure var.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Fixes a failure on masters if you explicitly set
openshift_docker_hosted_registry_insecure=true. This is the default but
if you tried to set it an error would trigger as a relevant variable was
not passed in the master playbooks.
Fixes setting the variable to false being ignored.
master/node playbooks were referencing the docker fact, which was not
set at that point and thus we were always getting the default of true,
regardless what was in your inventory.
Stop passing registry insecure in via playbooks, we can access it when running
openshift_facts itself. Add a new default in openshift facts.
|
|/ / |
|
|/ |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
If doing an upgrade with a 3.2.0.x version that was older than the
latest, the upgrade would actually use the latest in the systemd unit
files and thus the actual containers that get used. (despite pulling
down the correct version first)
|
|
|
|
|
|
| |
QE found that for fresh installs we were basing the docker version facts of the
images that could be pulled prior to configuring /etc/sysconfig/docker. This
is an edge case but something we need to fix.
|
|
|
|
| |
containerized env
|
| |
|
| |
|
|
|
|
| |
and atomic-openshift-master-controllers
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Currently there's no good way to install from a registry that requires
authentication. This applies both to RPM and containerized installs:
https://bugzilla.redhat.com/show_bug.cgi?id=1316341
The workaround is to 'docker login' as root and then have ansible pull the
images to the image cache.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is the containerized openshift_pkg_version equivalent. Originally I was
hoping to reuse openshift_pkg_version for containerized installs but the fact
that it's very coupled to yum made that pretty ugly.
However, I did opt to rely on the previously existing 'openshift_version'
variable. Containerized and RPM installs can both use that variable and it
will be set appropriately if either openshift_pkg_version or
openshift_image_tag are set. I suspect someday containerized installs will be
the only option and I didn't can to have thinkgs like openshift_pkg_version and
openshift_image_tag in the playbooks anymore the necessary.
|
| |
|
| |
|
|
|