diff --git a/.github/workflows/canary-integration-test.yml b/.github/workflows/canary-integration-test.yml
index e92f79f35fd74..d5f8a26346ab4 100644
--- a/.github/workflows/canary-integration-test.yml
+++ b/.github/workflows/canary-integration-test.yml
@@ -953,6 +953,23 @@ jobs:
     - name: write an object to one cluster, read from the other
       run: tests/scripts/github-action-helper.sh write_object_to_cluster1_read_from_cluster2
 
+    - name: verify the RGW configuration period was committed for both clusters
+      run: |
+        committed_msg="committing changes to RGW configuration period for CephObjectStore"
+        tests/scripts/github-action-helper.sh verify_operator_log_message "${committed_msg}" rook-ceph
+        tests/scripts/github-action-helper.sh verify_operator_log_message "${committed_msg}" rook-ceph-secondary
+
+    - name: restart operators for both clusters
+      run: |
+        tests/scripts/github-action-helper.sh restart_operator rook-ceph
+        tests/scripts/github-action-helper.sh restart_operator rook-ceph-secondary
+
+    - name: wait for both clusters to report that the RGW configuration period was NOT committed on the second reconcile
+      run: |
+        not_committed_msg="there are no changes to commit for RGW configuration period for CephObjectStore"
+        tests/scripts/github-action-helper.sh wait_for_operator_log_message "${not_committed_msg}" 60 rook-ceph
+        tests/scripts/github-action-helper.sh wait_for_operator_log_message "${not_committed_msg}" 20 rook-ceph-secondary
+
     - name: upload test result
       uses: actions/upload-artifact@v2
       if: always()
diff --git a/pkg/operator/ceph/object/admin.go b/pkg/operator/ceph/object/admin.go
index bac57005ce118..5fedaa51c29d0 100644
--- a/pkg/operator/ceph/object/admin.go
+++ b/pkg/operator/ceph/object/admin.go
@@ -315,11 +315,14 @@ func CommitConfigChanges(c *Context) error {
 		return errors.Wrap(err, "failed to determine if the staged RGW configuration period is different from current")
 	}
 
+	nsName := fmt.Sprintf("%s/%s", c.clusterInfo.Namespace, c.Name)
 	if !shouldCommit {
-		nsName := fmt.Sprintf("%s/%s", c.clusterInfo.Namespace, c.Name)
-		logger.Debugf("not committing changes to RGW configuration period for CephObjectStore %q", nsName)
+		// DO NOT MODIFY THE MESSAGE BELOW. It is checked in integration tests.
+		logger.Infof("there are no changes to commit for RGW configuration period for CephObjectStore %q", nsName)
 		return nil
 	}
+	// DO NOT MODIFY THE MESSAGE BELOW. It is checked in integration tests.
+	logger.Infof("committing changes to RGW configuration period for CephObjectStore %q", nsName)
 	// don't expect json output since we don't intend to use the output from the command
 	_, err = runAdminCommand(c, false, "period", "update", "--commit")
 	if err != nil {
diff --git a/tests/scripts/github-action-helper.sh b/tests/scripts/github-action-helper.sh
index 3f5d913c0e24c..2e5fe2616d7c9 100755
--- a/tests/scripts/github-action-helper.sh
+++ b/tests/scripts/github-action-helper.sh
@@ -14,7 +14,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-set -xe
+set -xeEo pipefail
 
 #############
 # VARIABLES #
@@ -237,6 +237,35 @@ function wait_for_rgw_pods() {
 
 }
 
+function verify_operator_log_message() {
+  local message="$1"  # param 1: the message to verify exists
+  local namespace="${2:-rook-ceph}"  # optional param 2: the namespace of the CephCluster (default: rook-ceph)
+  kubectl --namespace "$namespace" logs deployment/rook-ceph-operator | grep "$message"
+}
+
+function wait_for_operator_log_message() {
+  local message="$1"  # param 1: the message to look for
+  local timeout="$2"  # param 2: the timeout for waiting for the message to exist
+  local namespace="${3:-rook-ceph}"  # optional param 3: the namespace of the CephCluster (default: rook-ceph)
+  start_time="${SECONDS}"
+  while [[ $(( SECONDS - start_time )) -lt $timeout ]]; do
+    if verify_operator_log_message "$message" "$namespace"; then return 0; fi
+    sleep 5
+  done
+  echo "timed out" >&2 && return 1
+}
+
+function restart_operator () {
+  local namespace="${1:-rook-ceph}"  # optional param 1: the namespace of the CephCluster (default: rook-ceph)
+  get_pod_cmd=(kubectl --namespace "$namespace" get pod --selector app=rook-ceph-operator --no-headers)
+  pod="$("${get_pod_cmd[@]}" --output name)"
+  # --output name gives "pod/<name>", so don't specify the kind here
+  kubectl --namespace "$namespace" delete "$pod" # waits for pod to be deleted but not new pod to be running
+  timeout 20 bash -c \
+    "until [[ -n \"\$(${get_pod_cmd[*]} --field-selector=status.phase=Running 2>/dev/null)\" ]] ; do echo waiting && sleep 1; done"
+  "${get_pod_cmd[@]}"
+}
+
 function deploy_second_rook_cluster() {
   BLOCK=$(sudo lsblk|awk '/14G/ {print $1}'| head -1)
   cd cluster/examples/kubernetes/ceph/
@@ -275,7 +304,5 @@ EOF
 FUNCTION="$1"
 shift # remove function arg now that we've recorded it
 # call the function with the remainder of the user-provided args
-if ! $FUNCTION "$@"; then
-  echo "Call to $FUNCTION was not successful" >&2
-  exit 1
-fi
+# -e, -E, and -o=pipefail will ensure this script returns a failure if a part of the function fails
+$FUNCTION "$@"