diff options
| author | Scott Dodson <sdodson@redhat.com> | 2017-08-01 12:55:47 -0400 |
|---|---|---|
| committer | Scott Dodson <sdodson@redhat.com> | 2017-08-22 12:57:57 -0400 |
| commit | 4b5d8d2dc25dbca20be59f3d5d111d737fd865bc (patch) | |
| tree | b2707928f22b8a308ff8b110b2ce466e0a9b3c95 /playbooks/common/openshift-etcd/scaleup.yml | |
| parent | b107f677a3563f6db9a1ef17fe10518d321d495c (diff) | |
| download | openshift-4b5d8d2dc25dbca20be59f3d5d111d737fd865bc.tar.gz openshift-4b5d8d2dc25dbca20be59f3d5d111d737fd865bc.tar.bz2 openshift-4b5d8d2dc25dbca20be59f3d5d111d737fd865bc.tar.xz openshift-4b5d8d2dc25dbca20be59f3d5d111d737fd865bc.zip | |
Switch to migrating one host and forming a new cluster
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
Diffstat (limited to 'playbooks/common/openshift-etcd/scaleup.yml')
| -rw-r--r-- | playbooks/common/openshift-etcd/scaleup.yml | 13 |
1 files changed, 13 insertions, 0 deletions
diff --git a/playbooks/common/openshift-etcd/scaleup.yml b/playbooks/common/openshift-etcd/scaleup.yml index 192305bc8..52b90daca 100644 --- a/playbooks/common/openshift-etcd/scaleup.yml +++ b/playbooks/common/openshift-etcd/scaleup.yml @@ -24,6 +24,9 @@ member add {{ etcd_hostname }} {{ etcd_peer_url_scheme }}://{{ etcd_ip }}:{{ etcd_peer_port }} delegate_to: "{{ etcd_ca_host }}" register: etcd_add_check + retries: 3 + delay: 10 + until: etcd_add_check.rc == 0 roles: - role: openshift_etcd when: etcd_add_check.rc == 0 @@ -36,3 +39,13 @@ r_etcd_common_etcd_runtime: "{{ openshift.common.etcd_runtime }}" - role: nickhammond.logrotate when: etcd_add_check.rc == 0 + post_tasks: + - name: Verify cluster is stable + command: > + /usr/bin/etcdctl --cert-file {{ etcd_peer_cert_file }} + --key-file {{ etcd_peer_key_file }} + --ca-file {{ etcd_peer_ca_file }} + -C {{ etcd_peer_url_scheme }}://{{ hostvars[etcd_ca_host].etcd_hostname }}:{{ etcd_client_port }} + cluster-health + retries: 1 + delay: 30 |