| Age | Commit message (Collapse) | Author |
|
This can be used in place of:
inventory_hostname in groups['oo_masters_to_config'] or \
inventory_hostname in groups['oo_nodes_to_config']
While the previous version works, this change helps with the tasks are
run during AMI creation.
Code written by Kenny Woodson @kwoodson.
|
|
cri-o: open port 10010
|
|
crio: Add failed_when to overlay check
|
|
Correct openshift_release regular expression
|
|
crio: restorcon /var/lib/containers
|
|
Force include_role to static for loading openshift_facts module
|
|
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
|
|
docker: set credentials when using system container
|
|
1504593 Refactor health check playbooks
|
|
Remove 'Not Started' status from playbook checkpoint
|
|
Fix CA Bundle passed to service-catalog broker for ansible-service-broker
|
|
Make openshift-ansible depend on all subpackages
|
|
Remove inadvertently committed inventory file
|
|
|
|
Change dnsmasq to bind-interfaces + except-interfaces
|
|
This previously passed all CI tests, merge queue broke last night.
Merging, this is a blocker bug.
|
|
Created by command:
/usr/bin/tito tag --debug --accept-auto-changelog --keep-version --debug
|
|
When using the cri-o system container some items in /var/lib/containers
end up having incorrect labels. This runs restorecon -R
/var/lib/containers after the system container is installed.
|
|
Match only versions like:
3.7
3.7.0
3.7.0.0
Avoids matches like:
3.7.0.0.0.0
|
|
Created by command:
/usr/bin/tito tag --debug --accept-auto-changelog --keep-version --debug
|
|
Moving removal of unwanted artifacts to image_prep.
|
|
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1506399
Signed-off-by: Steve Milner <smilner@redhat.com>
|
|
Documentation fix for openshift_aws_iam_cert_key_path
|
|
Ensure journald persistence directories exist
|
|
|
|
Automatic merge from submit-queue.
Openshift MGMT Fixes and Container Provider Integration
Various small fixes/polishing to the openshift management role.
* Updated uninstall script
* Automatically add cluster to new management deployment
* Scaffolding for scripts to add *N* number of OCP/Origin clusters as Container Providers in any arbitrary MIQ/CFME instance
* Fix mis-named task in `openshift_nfs/create_export`
* Uninstall playbook only runs on one master
|
|
Created by command:
/usr/bin/tito tag --debug --accept-auto-changelog --keep-version --debug
|
|
Fix typo and add detailed comments in kuryr
|
|
Automatic merge from submit-queue.
Fix loop_var warnings during logging install
Currently, install_logging.yml in openshift_logging role
loops over other roles.
This creates a collision with the keyword 'item' in those roles.
This commit adds the loop_var as suggested by ansible warnings.
|
|
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
|
|
Created by command:
/usr/bin/tito tag --debug --accept-auto-changelog --keep-version --debug
|
|
Automatic merge from submit-queue.
etcd: remove hacks for the system container
This removes the special handling of the etcd directories when running as a system container.
Marked as WIP as it depends from this [change](https://bugzilla.redhat.com/show_bug.cgi?id=1484326) in the etcd image to be reflected into `registry.access.redhat.com/rhel7/etcd`
|
|
Currently, we have to set the listen ip for dnsmasq via ansible
during installation.
This commit enables dnsmasq to bind-interfaces + exclude interfaces
to ensure dnsmasq doesn't listen on lo interface.
|
|
Automatic merge from submit-queue.
Ensure deployment_subtype is set within openshift_sanitize_inventory.
https://bugzilla.redhat.com/show_bug.cgi?id=1504021
|
|
1505537 Remove pause from master service startup
|
|
Remove base package install
|
|
|
|
|
|
Automatic merge from submit-queue.
Remove unused registry_volume_claim variable
The "registry_volume_claim" variable has not been used since at least
commit 7cf5cc1 (February 21, 2017) and this commit removes the last
mention.
|
|
Currently, we configure openshfit masters to modify journald
to use persistent storage. The directory structure must be
created manually according to documentation.
This commit ensures the needed directory is created.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1506141
|
|
Sync latest imagestreams and templates
|
|
Automatic merge from submit-queue.
Add installer checkpoint for prometheus
|
|
Fixes 1505537
https://bugzilla.redhat.com/show_bug.cgi?id=1505537
|
|
Change default in prometheus storage type to emptydir
|
|
Currently, in 1.5, 3.6, 1.7 upgrade-all plays, control
plane upgrades are not called correctly.
This commit ensures the master config hook is appropriately
applied during these upgrades to match the steps in
control plane only upgrades.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1486054
|
|
Playbook status should only be reported for phases that have completed
or which have been started. Reporting 'Not Started' could be confusing
when running playbooks that do not have an 'installer phase'.
Fixes #5675
|
|
Automatic merge from submit-queue.
Bug 1491636 - honor node selectors
The deprecation of `*_hosted_*` vars made logging node selectors set in the inventory to be ignored.
Node selectors were set as 'facts' and they have higher priority than inventory variables. Setting logging node selectors could be achieved only by using command line --extra-vars.
It is related to the same BZ as https://github.com/openshift/openshift-ansible/pull/5858 but the issue is different.
|
|
Created by command:
/usr/bin/tito tag --debug --accept-auto-changelog --keep-version --debug
|
|
Automatic merge from submit-queue.
Split prometheus image defaults to prefix and version
To allow easier testing and deployment from different repos, image defaults were split to prefix and version that can be set externally.
bz: https://bugzilla.redhat.com/show_bug.cgi?id=1493431
|
|
smarterclayton/allow_per_node_group_bootstrap_and_image
Automatic merge from submit-queue.
Handle bootstrap behavior in GCP template
Allow each node group to request bootstrap, allow per node group image
override, and ensure the provision logic does not wait for bootstrapping
node groups before continuing.
This is an incremental step to allow GCP clusters to use bootstrap logic
on cluster deploy without having fully baked images. We will switch over
slowly and ensure both code paths function. Then we can remove this as
necessary.
For metadata, we set the cluster id and bootstrap state into instance metadata. On GCP, we'll use project metadata to set the bootstrap kubeconfig file and a startup-script to call it (not in a PR yet).
Pairs with openshift/origin-gce#54 @kwoodson
|