From a6d5c625956a5051b7bbd9fc48430b9df11084ee Mon Sep 17 00:00:00 2001 From: Michael Gugino Date: Fri, 8 Dec 2017 18:30:09 -0500 Subject: Refactor node upgrade to include less serial tasks This commit moves the pulling of images, packages, and updating config files into a non-serialized play. The serialized play is now in charge of marking unschedulable, draining, stopping and restarting services, and marking schedulable. If rpm install / container download takes 60s per host, this will save 3 hours and 10 minutes at 200 hosts per cluster and forks of 20 hosts. --- .../openshift-cluster/upgrades/upgrade_nodes.yml | 33 ++++++++++++++-------- 1 file changed, 22 insertions(+), 11 deletions(-) (limited to 'playbooks/common') diff --git a/playbooks/common/openshift-cluster/upgrades/upgrade_nodes.yml b/playbooks/common/openshift-cluster/upgrades/upgrade_nodes.yml index f7a85545b..a3cb1d0f9 100644 --- a/playbooks/common/openshift-cluster/upgrades/upgrade_nodes.yml +++ b/playbooks/common/openshift-cluster/upgrades/upgrade_nodes.yml @@ -1,16 +1,25 @@ --- +- name: Prepull images and rpms before doing rolling restart + hosts: oo_nodes_to_upgrade:!oo_masters_to_config + roles: + - role: openshift_facts + tasks: + - include_role: + name: openshift_node + tasks_from: upgrade_pre.yml + vars: + openshift_node_upgrade_in_progress: True + - name: Drain and upgrade nodes hosts: oo_nodes_to_upgrade:!oo_masters_to_config # This var must be set with -e on invocation, as it is not a per-host inventory var # and is evaluated early. Values such as "20%" can also be used. serial: "{{ openshift_upgrade_nodes_serial | default(1) }}" max_fail_percentage: "{{ openshift_upgrade_nodes_max_fail_percentage | default(0) }}" - + roles: + - lib_openshift + - openshift_facts pre_tasks: - - name: Load lib_openshift modules - import_role: - name: lib_openshift - # TODO: To better handle re-trying failed upgrades, it would be nice to check if the node # or docker actually needs an upgrade before proceeding. Perhaps best to save this until # we merge upgrade functionality into the base roles and a normal config.yml playbook run. @@ -33,18 +42,12 @@ retries: 60 delay: 60 - roles: - - openshift_facts post_tasks: - include_role: name: openshift_node tasks_from: upgrade.yml vars: openshift_node_upgrade_in_progress: True - - include_role: - name: openshift_excluder - vars: - r_openshift_excluder_action: enable - name: Set node schedulability oc_adm_manage_node: node: "{{ openshift.node.nodename | lower }}" @@ -55,3 +58,11 @@ register: node_schedulable until: node_schedulable|succeeded when: node_unschedulable|changed + +- name: Re-enable excluders + hosts: oo_nodes_to_upgrade:!oo_masters_to_config + tasks: + - include_role: + name: openshift_excluder + vars: + r_openshift_excluder_action: enable -- cgit v1.2.3