fix: site replication healing of missing buckets (#15298)

fixes a regression from #15186

- Adding tests to cover healing of buckets.
- Also dereference quota in SiteReplicationStatus only when non-nil

docs/site-replication/run-multi-site-minio-idp.sh

@@ -317,7 +317,8 @@ fi
 kill -9 ${site1_pid}
 # Update tag on minio2/newbucket when minio1 is down
 ./mc tag set minio2/newbucket "key=val2"
-
+# create a new bucket on minio2. This should replicate to minio1 after it comes online.
+./mc mb minio2/newbucket2
 # Restart minio1 instance
 minio server --config-dir /tmp/minio-internal --address ":9001" /tmp/minio-internal-idp1/{1...4} >/tmp/minio1_1.log 2>&1 &
 sleep 15
@@ -328,3 +329,9 @@ if [ "${val}" != "val2" ]; then
     echo "expected bucket tag to have replicated, exiting..."
     exit_1;
 fi
+# Test if bucket created when minio1 is down healed
+diff -q <(./mc ls minio1 | awk '{print $3}') <(./mc ls minio2 | awk '{print $3}') 1>/dev/null
+if  [ $? -ne 0 ]; then
+    echo "expected bucket to have replicated, exiting..."
+    exit_1;
+fi