Make C/R work with nodes > 1. This fix makes sure that the app coordinators send

the "ready-to-checkpoint" signal to the global coordinator only after ORTE has
initialized.

This commit was SVN r26795.

orte/mca/ess/env/ess_env_module.c

@@ -336,9 +336,7 @@ static int rte_ft_event(int state)
                 exit_status = ret;
                 goto cleanup;
             }
-            while (coll.active) {
-                opal_progress();
-            }
+            ORTE_WAIT_FOR_COMPLETION(coll.active);
 
             if( orte_cr_flush_restart_files ) {
                 OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
@@ -463,9 +461,7 @@ static int rte_ft_event(int state)
             exit_status = ret;
             goto cleanup;
         }
-        while (coll.active) {
-            opal_progress();
-        }
+	ORTE_WAIT_FOR_COMPLETION(coll.active);
 
         if( orte_cr_flush_restart_files ) {
             OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,

orte/mca/snapc/full/snapc_full_module.c

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2004-2010 The Trustees of Indiana University.
+ * Copyright (c) 2004-2012 The Trustees of Indiana University.
  *                         All rights reserved.
  * Copyright (c) 2004-2011 The Trustees of the University of Tennessee.
  *                         All rights reserved.
@@ -194,6 +194,13 @@ int orte_snapc_full_module_init(bool seed, bool app)
      * Application Snapshot Coordinator
      */
     else if(app) {
+        /*
+         * Start the app coordinator only after ORTE has initialized.
+         */
+        if(!orte_initialized) {
+            goto cleanup;
+        }
+
         opal_output_verbose(5, mca_snapc_full_component.super.output_handle,
                             "snapc:full: module_init: Application Snapshot Coordinator");