ed3b27cc92
This commit addresses a few shortcomings in the etcd service:
* Adding or removing etcd nodes required manual intervention.
* The etcd service would have brief outages during upgrades or
reconfigures because restarts weren't always serialised.
This makes the etcd service follow a similar pattern to mariadb:
* There is now a distiction between bootstrapping the cluster
and adding / removing another member.
* This more closely follows etcd's upstream bootstrapping
guidelines.
* The etcd role now serialises restarts internally so the
kolla_serial pattern is no longer appropriate (or necessary).
This does not remove the need for manual intervention in all
failure modes: the documentation has been updated to address the
most common issues.
Note that there's repetition in the container specifications: this
is somewhat deliberate. In a future cleanup, it's intended to reduce
the duplication.
Change-Id: I39829ba0c5894f8e549f9b83b416e6db4fafd96f
27 lines
867 B
YAML
27 lines
867 B
YAML
---
|
|
- name: Ensure etcd volume
|
|
become: true
|
|
kolla_container:
|
|
action: "create_volume"
|
|
common_options: "{{ docker_common_options }}"
|
|
name: "kolla_etcd"
|
|
register: etcd_volume
|
|
|
|
# NOTE(jan.gutter): If the play is interrupted before properly bootstrapping,
|
|
# we will incorrectly assume that an etcd cluster exists. This likely requires
|
|
# manual intervention to unwedge. If a volume exists we must assume there's
|
|
# data on it.
|
|
|
|
- name: Divide hosts by their etcd volume availability
|
|
group_by:
|
|
key: etcd_had_volume_{{ etcd_volume is not changed }}
|
|
changed_when: false
|
|
|
|
- name: Establish whether the cluster has already existed
|
|
set_fact:
|
|
etcd_cluster_exists: "{{ groups.etcd_had_volume_True is defined }}"
|
|
changed_when:
|
|
- etcd_remove_deleted_members | bool
|
|
- groups.etcd_had_volume_True is defined
|
|
notify: Check for deleted members
|