From: Giulio Fidente <gfidente@redhat.com>
Date: Tue, 18 Jul 2017 09:03:35 +0000 (+0200)
Subject: Use optimal (instead of default) tunables for Ceph on upgrade
X-Git-Tag: opnfv-6.0.0~384^2
X-Git-Url: https://gerrit.opnfv.org/gerrit/gitweb?a=commitdiff_plain;h=5e9f855f7c96950ca29a0f85086441c57ae7aed5;p=apex-tripleo-heat-templates.git

Use optimal (instead of default) tunables for Ceph on upgrade

With the default setting, after the majority of the monitors have
been upgraded the cluster will go in WARN state because of legacy
tunables. This changes the tunables we set after each monitor is
upgraded from 'default' to 'optimal' [1].

1. http://docs.ceph.com/docs/master/rados/operations/crush-map/#warning-when-tunables-are-non-optimal

Change-Id: I0f16c29cc200d762f0c4acfd87ba7d1adb5c1eeb
Closes-Bug: #1704959
---

diff --git a/puppet/services/ceph-mon.yaml b/puppet/services/ceph-mon.yaml
index 4fe6e908..3e4f5b42 100644
--- a/puppet/services/ceph-mon.yaml
+++ b/puppet/services/ceph-mon.yaml
@@ -172,6 +172,6 @@ outputs:
           until: ceph_quorum_nodecheck.rc == 0
           retries: {get_param: CephValidationRetries}
           delay: {get_param: CephValidationDelay}
-        - name: ceph osd crush tunables default
+        - name: set crush tunables
           tags: step0
-          shell: ceph osd crush tunables default
+          shell: ceph osd crush tunables optimal