From 9eacde212f7a322b3c61b8ef1074435c1d02f4bd Mon Sep 17 00:00:00 2001
From: Florian Ruynat <16313165+floryut@users.noreply.github.com>
Date: Wed, 27 Oct 2021 00:23:09 +0200
Subject: [PATCH] Fix quorum check when recovering broken etcd cluster (#8126)

---
 roles/etcd/tasks/join_etcd_member.yml           | 3 +++
 roles/recover_control_plane/etcd/tasks/main.yml | 3 +--
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/roles/etcd/tasks/join_etcd_member.yml b/roles/etcd/tasks/join_etcd_member.yml
index 28d259ccc..22440394f 100644
--- a/roles/etcd/tasks/join_etcd_member.yml
+++ b/roles/etcd/tasks/join_etcd_member.yml
@@ -32,6 +32,9 @@
   register: etcd_member_in_cluster
   changed_when: false
   check_mode: no
+  retries: "{{ etcd_retries }}"
+  delay: "{{ retry_stagger | random + 3 }}"
+  until: etcd_member_in_cluster.rc == 0
   tags:
     - facts
   environment:
diff --git a/roles/recover_control_plane/etcd/tasks/main.yml b/roles/recover_control_plane/etcd/tasks/main.yml
index e3dc33930..45e2c65e4 100644
--- a/roles/recover_control_plane/etcd/tasks/main.yml
+++ b/roles/recover_control_plane/etcd/tasks/main.yml
@@ -20,10 +20,9 @@
   when:
     - groups['broken_etcd']
 
-# When there is an error, everything is printed in stderr_lines, even "is healthy" messages.
 - name: Set has_quorum fact
   set_fact:
-    has_quorum: "{{ etcd_endpoint_health.stderr_lines | select('match', '.*is healthy.*') | list | length >= etcd_endpoint_health.stderr_lines | select('match', '.*is unhealthy.*') | list | length }}"
+    has_quorum: "{{ etcd_endpoint_health.stdout_lines | select('match', '.*is healthy.*') | list | length >= etcd_endpoint_health.stderr_lines | select('match', '.*is unhealthy.*') | list | length }}"
   when:
     - groups['broken_etcd']