Enable the system to keep functioning even when multiple launches are occurring simultaneously.

This is a bit of a hack, but it does seem to allow the system to work. A better solution is being discussed.

This commit was SVN r21705.

orte/mca/plm/base/plm_base_launch_support.c

@@ -980,7 +980,8 @@ int orte_plm_base_report_launched(orte_jobid_t job)
     ORTE_PROGRESSED_WAIT(app_launch_failed, jdata->num_launched, jdata->num_procs);
 
     /* cancel the lingering recv */
-    if (ORTE_SUCCESS != (rc = orte_rml.recv_cancel(ORTE_NAME_WILDCARD, ORTE_RML_TAG_APP_LAUNCH_CALLBACK))) {
+    if (ORTE_SUCCESS != (rc = orte_rml.recv_cancel(ORTE_NAME_WILDCARD, ORTE_RML_TAG_APP_LAUNCH_CALLBACK)) &&
+        ORTE_ERR_NOT_FOUND != rc) {
         ORTE_ERROR_LOG(rc);
         return rc;
     }

orte/mca/plm/base/plm_base_receive.c

@@ -381,6 +381,7 @@ void orte_plm_base_recv(int status, orte_process_name_t* sender,
                                                       NULL))) {
         ORTE_ERROR_LOG(rc);
     }
+    
     return;
 }
 

orte/mca/plm/base/plm_base_rsh_support.c

@@ -1394,8 +1394,6 @@ void orte_plm_base_reset_job(orte_job_t *jdata)
             }
             /* adjust job accounting */
             jdata->num_terminated--;
-            jdata->num_launched--;
-            jdata->num_reported--;
         }
     }
     /* clear the info on who aborted */
@@ -1404,4 +1402,7 @@ void orte_plm_base_reset_job(orte_job_t *jdata)
         OBJ_RELEASE(jdata->aborted_proc);  /* maintain reference count */
         jdata->aborted_proc = NULL;
     }
+    /* since every daemon will be reporting status for every proc, reset these to zero */
+    jdata->num_launched = 0;
+    jdata->num_reported = 0;
 }