| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Automatic merge from submit-queue.
Migrate embedded etcd to external etcd
Trello card: https://trello.com/c/9fnBfkT5/528-migrate-embedded-etcd-hosts-to-external-process?menu=filter&filter=label:committed-3.7
Tested on:
- [x] embedded etcd -> external rpm etcd
- [x] embedded etcd -> external docker etcd
One needs to set the `openshift_image_tag` before running the migration to containerized docker etcd.
|
| | |
|
| | |
|
|\ \
| | |
| | |
| | |
| | |
| | |
| | | |
Automatic merge from submit-queue.
Allow checkpoint status to work across all groups
Conditionals placed in inventories were not being applied to localhost causing the checkpoint status to not be updated properly. Moving to the `oo_all_hosts` group will correctly pick up the conditional and apply regardless of which group (or host) may have that conditional set.
|
| |/ |
|
|\ \
| |/
|/|
| |
| |
| |
| |
| |
| | |
Automatic merge from submit-queue.
Refactor certificate playbooks for checkpointing
The recent breakout of the certificate playbooks resulted in the certificate portion of the install to be outside of the component checkpoint. This PR moves the certificate playbooks inside their respective config.yml playbook.
Removes an unused tasks file from the openshift_named_certificates role.
|
| | |
|
|\ \
| |/
|/|
| |
| |
| |
| |
| |
| | |
Automatic merge from submit-queue.
Use etcd_ip when communicating with the cluster as a peer in etcd scaleup.
We can't communicate with the hostname here since etcd will be [configured to listen to its IP address](https://github.com/openshift/openshift-ansible/blob/ead22bb1c3b6d6366502b14b97b7aae5605f8a58/roles/etcd/defaults/main.yaml#L69).
Fixes https://github.com/openshift/openshift-ansible/issues/5620
|
| | |
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
| |
- Added installer_checkpoint role and callback plugin
- Added checkpoint 'Start' and 'End' plays to each installation phase
Additional items related to enabling proper checkpointing:
- Removed openshift_set_hostname and related task (related to 3.0)
- Added openshift-hosted entry point playbook
- Moved openshift metrics and logging out of openshift_hosted playbook
- Moved API Aggregation play to master install
|
|\
| |
| |
| |
| |
| |
| | |
Automatic merge from submit-queue
Consolidate etcd common role
The last PR toward a single etcd role
|
| | |
|
|/ |
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Automatic merge from submit-queue
Improve CA redeploy restart logic
Expired etcd certificates require special casing around restarts in the certificate redeploy playbooks. When etcd certificates are expired we can't restart masters or nodes. We also can't simply restart etcd because peers also had expired certificates so we must start/stop etcd when we detect expired etcd certificates.
`openshift-ca.yml`:
* No longer restart master services when etcd certificates were previously expired.
* No longer restart node services when master or etcd certificates were previously expired.
`etcd-ca.yml`:
* No longer restart master services when etcd certificates were previously expired.
Tested using [gen_expired_tls.sh](https://gist.github.com/abutcher/bdd20b9d582675d89fb22658689c49e4) on one of my master/etcd hosts to ensure that restart logic changes caused us to skip the right restarts and do a full start/stop of etcd in the `redeploy-certificates.yml` and `redeploy-etcd-certificates.yml` playbooks.
Note: When this happens with a cluster and you want to replace all certificates, you can run: `redeploy-etcd-ca.yml`, `redeploy-openshift-ca.yml` (which will both skip restarts) and then run `redeploy-certificates.yml` which will now be able to full/stop start etcd.
|
| | |
|
|\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Automatic merge from submit-queue
Set master facts prior to adding new etcd client urls to master config.
`openshift_master` role dependencies were moved out of the role in https://github.com/openshift/openshift-ansible/pull/5392 so we need to call `openshift_master_facts` prior to patching the master config in etcd scaleup.
https://bugzilla.redhat.com/show_bug.cgi?id=1490304
|
| |/ |
|
|/ |
|
|\
| |
| | |
Use openshift.common.ip rather than ansible_default_ipv4 in etcd migration playbook.
|
| |
| |
| |
| | |
playbook.
|
|/ |
|
| |
|
|\
| |
| | |
Move firewall install and fix scaleup playbooks
|
| | |
|
|\ \
| |/
|/| |
Fix scaleup on containerized installations
|
| |
| |
| |
| |
| |
| | |
The output from member add had quotes added that caused etcd to fail.
Also there were some places that needed to have retries added because
containerized start/stop/restart are slower than rpm based installs.
|
|/ |
|
|
|
|
|
|
|
|
| |
With large datasets where there are many keys with TTLs the expiry was
creating a data inconsistency problem. The hope is that by performing
the migration once and then forming a new cluster this is avoided.
Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1475351
|
|
|
|
| |
new member registration.
|
|\
| |
| | |
Add etcd scaleup playbook
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| | |
- move openshift-etcd/migrate.yml from byo to common and keep just the entry point
- replace std_include with essential plays (e.g. no need to detect openshift version)
- delegate the ttl re-attaching to the first master
|
|/ |
|
| |
|
| |
|
|
|
|
|
|
|
| |
The add_host: task does not change any data on the host and as practice
has been configured to changed_when: False. This commit standardizes
that usage in the byo and common playbooks. Additionally, taks names
are added to each task to improve troubleshooting.
|
|
|
|
|
| |
* Added checks to make ci for yaml linting
* Modified y(a)ml files to pass lint checks
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
certificates.
|
|
|
|
|
|
| |
- gather facts requiring docker only if docker is present and running
- Update reference to etcd role in playbooks/common/openshift-etcd/config.yml
to use openshift_etcd
|
|
|
|
|
|
|
| |
- Prevents roles that need common facts from needing to require
openshift_common, which pulls in the openshift binary.
- Add dependency on openshift_facts to os_firewall, since it uses
openshift.common facts
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- refactors the docker role to push generic config into docker role and wrap
openshift specific variables into an openshift_docker role and it's
dependent openshift_docker_facts role
- adds support for setting --confirm-def-push flag (Resolves
https://github.com/openshift/openshift-ansible/issues/1014)
- moves docker related facts from common/node roles to a new docker role
- renames cli_docker_* role varialbes to openshift_docker-* (maintaining
backward compatibility)
- update role dependencies to pull in openshift_docker conditionally based on
is_containerized
- remove playbooks/common/openshift-docker since the docker role is now
conditionally included
|
|\
| |
| | |
Install and start one etcd server before the others.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In faster environments (i.e. all local VMs) etcd nodes could come online at
roughly the same time leading to conflicts with self-elections, resulting in a
non-functional cluster.
To solve we configure the first etcd host by itself, then configure the
remaining ones in parallel to keep things as fast as possible.
|