| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Keeping the dynamic inventory in the `sample-inventory` alongside the
`group_vars` poses a problem: when we update the inventory, our users
won't get the latest version.
The dynamic inventory should be independent of the OpenShift or
OpenStack configuration and the users should be able to either pass it
explicitly or create a symlink if they want.
|
|
When deploying on OpenStack with internal DNS configured, this will set
`openshift_hostname` to the Nova server name instead of its IP address.
Without those two matching, the OpenStack cloud provider configuration
will fail and the OpenShift nodes will not start.
|
|
As described in[1], OpenShift currently only works with Block Storage
API v2 and the version autodetection is failing to figure that out.
[1]: https://github.com/openshift/openshift-docs/issues/5730
|
|
Automatic merge from submit-queue.
Openstack fixes
This includes a few fixes for the OpenStack provider.
It should fix #6555 and possibly also #6560.
|
|
The OpenStack dynamic inventory was setting the
`openshift_node_labels` value as a string which causes a failure with
the `lib_utils_oo_dict_to_keqv_list` filter.
Fixes #6555
|
|
|
|
|
|
Optionally subscribe OpenStack RHEL nodes
|
|
openshift.master.cluster_method.
With pacemaker removed there is no longer a need for
openshift.master.cluster_method. We only have one option.
|
|
Signed-off-by: Bogdan Dobrelya <bdobreli@redhat.com>
|
|
Use testing openshift repos to sidestep packaging issues for 3.6 clients being
used in 3.7 deployments or something.
Poke host requirements Centos 7.4 as 7.3 can't see that testing repo.
Signed-off-by: Bogdan Dobrelya <bdobreli@redhat.com>
|
|
|
|
|
|
Unlike other cloud providers, OpenStack VMs are not able to resolve each
other by their names. If you try to run the playbooks against nodes
without a pre-created /etc/hosts or a DNS that provides the hostname/ip
resolution, it will fail.
By setting the `openshift_hostname` variable to each node's IP address,
we're able to deploy a functional cluster without running a custom DNS.
It is still possible to provide an external server with nsupdate keys
and have it be populated, but that is no longer a hard requirement.
|
|
|
|
This makes sure that all the variables used in the `openshift_openstack`
role are prefixed with `openshift_openstack_` as is the convention.
|
|
Most of the vars in `roles/openshift_openstack/defaults/main.yml` are
now prefixed with `openstack_`.
|
|
These options will have no effect until we add static inventory and
bastion support back in.
|
|
All the tasks that were previously in playbooks are now under
`roles/openshift_openstack`.
The `openshift-cluster` directory now only contains playbooks that
include tasks from that role. This makes the structure much closer to
that of the AWS provider.
|
|
We move them from `playbooks/provisioning/openstack` to
`playbooks/openstack` to mirror `playbooks/aws`.
|
