From d4f253ef9fea8dfdb4f7a27098ac28df8bef6a4e Mon Sep 17 00:00:00 2001
From: "Parsons, Cliff (cp769u)" <cp769u@att.com>
Date: Tue, 11 May 2021 15:31:03 +0000
Subject: [PATCH] Make Ceph pool init job consistent with helm test

The current pool init job only allows the finding of PGs in the
"peering" or "activating" (or active) states, but it should also
allow the other possible states that can occur while the PG
autoscaler is running ("unknown" and "creating" and "recover").
The helm test is already allowing these states, so the pool init
job is being changed to also allow them to be consistent.

Change-Id: Ib2c19a459c6a30988e3348f8d073413ed687f98b
---
 ceph-client/Chart.yaml                      | 2 +-
 ceph-client/templates/bin/pool/_init.sh.tpl | 6 +++---
 releasenotes/notes/ceph-client.yaml         | 1 +
 3 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/ceph-client/Chart.yaml b/ceph-client/Chart.yaml
index b2b053501..3c7363c9b 100644
--- a/ceph-client/Chart.yaml
+++ b/ceph-client/Chart.yaml
@@ -15,6 +15,6 @@ apiVersion: v1
 appVersion: v1.0.0
 description: OpenStack-Helm Ceph Client
 name: ceph-client
-version: 0.1.15
+version: 0.1.16
 home: https://github.com/ceph/ceph-client
 ...
diff --git a/ceph-client/templates/bin/pool/_init.sh.tpl b/ceph-client/templates/bin/pool/_init.sh.tpl
index 70a77191b..ec81a96c0 100644
--- a/ceph-client/templates/bin/pool/_init.sh.tpl
+++ b/ceph-client/templates/bin/pool/_init.sh.tpl
@@ -42,9 +42,9 @@ function wait_for_pgs () {
   # Loop until all pgs are active
   while [[ $pgs_ready -lt 3 ]]; do
     pgs_state=$(ceph --cluster ${CLUSTER} pg ls -f json | jq -c "${query}")
-    if [[ $(jq -c '. | select(.state | contains("peer") or contains("activating") | not)' <<< "${pgs_state}") ]]; then
-      # If inactive PGs aren't peering, fail
-      echo "Failure, found inactive PGs that aren't peering"
+    if [[ $(jq -c '. | select(.state | contains("peer") or contains("activating") or contains("recover") or contains("unknown") or contains("creating") | not)' <<< "${pgs_state}") ]]; then
+      # If inactive PGs aren't in the allowed set of states above, fail
+      echo "Failure, found inactive PGs that aren't in the allowed set of states"
       exit 1
     fi
     if [[ "${pgs_state}" ]]; then
diff --git a/releasenotes/notes/ceph-client.yaml b/releasenotes/notes/ceph-client.yaml
index 1ac93bf7f..fb36e57f5 100644
--- a/releasenotes/notes/ceph-client.yaml
+++ b/releasenotes/notes/ceph-client.yaml
@@ -16,4 +16,5 @@ ceph-client:
   - 0.1.13 Fix ceph-client helm test
   - 0.1.14 Allow Ceph RBD pool job to leave failed pods
   - 0.1.15 Make ceph-client helm test more PG specific
+  - 0.1.16 Make Ceph pool init job consistent with helm test
 ...