Bug #25741 Unable to allocate node id during restarting data node

- add additional sleep so that we don't get 100% cpu usage retrying saomething that can take some time before succeeding

storage/ndb/src/mgmsrv/MgmtSrvr.cpp

@@ -2167,6 +2167,11 @@ MgmtSrvr::alloc_node_id_req(NodeId free_node_id, enum ndb_mgm_node_type type)
         nodeId = refToNode(ref->masterRef);
 	if (!theFacade->get_node_alive(nodeId))
 	  nodeId = 0;
+        if (ref->errorCode != AllocNodeIdRef::NotMaster)
+        {
+          /* sleep for a while (100ms) before retrying */
+          NdbSleep_MilliSleep(100);  
+        }
         continue;
       }
       return ref->errorCode;