From bf8347e1491926418923bba9870109dfc3fbe89a Mon Sep 17 00:00:00 2001
From: Christian Zunker <christian.zunker@codecentric.de>
Date: Mon, 12 Jun 2017 08:30:22 +0000
Subject: [PATCH] Restart all OSDs and do not stop after the first one.

The current handler only restarts one OSD on each OSD server. After
the first one the handler stops, not matter what results the checks had.

Co-Authored-By: Gaudenz Steinlin (@gaudenz)
---
 roles/ceph-common/templates/restart_osd_daemon.sh.j2 | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/roles/ceph-common/templates/restart_osd_daemon.sh.j2 b/roles/ceph-common/templates/restart_osd_daemon.sh.j2
index 5d1d04597..f027574cf 100644
--- a/roles/ceph-common/templates/restart_osd_daemon.sh.j2
+++ b/roles/ceph-common/templates/restart_osd_daemon.sh.j2
@@ -8,7 +8,7 @@ check_pgs() {
   while [ $RETRIES -ne 0 ]; do
     test "[""$(ceph $CEPH_CLI -s -f json | python -c 'import sys, json; print(json.load(sys.stdin)["pgmap"]["num_pgs"])')""]" == "$(ceph $CEPH_CLI -s -f json | python -c 'import sys, json; print [ i["count"] for i in json.load(sys.stdin)["pgmap"]["pgs_by_state"] if i["state_name"] == "active+clean"]')"
     RET=$?
-    test $RET -eq 0 && exit 0
+    test $RET -eq 0 && return 0
     sleep $DELAY
     let RETRIES=RETRIES-1
   done
@@ -28,7 +28,7 @@ for id in $(ls /var/lib/ceph/osd/ | sed 's/.*-//'); do
   # Wait and ensure the socket exists after restarting the daemon
   SOCKET=/var/run/ceph/{{ cluster }}-osd.${id}.asok
   while [ $COUNT -ne 0 ]; do
-    test -S $SOCKET && check_pgs
+    test -S $SOCKET && check_pgs && continue 2
     sleep 1
     let COUNT=COUNT-1
   done