Added an option for daemonizing orted. The existing behavior to --no-daemonize

for gridengine is not changed.

This commit was SVN r15050.

orte/mca/pls/gridengine/pls_gridengine.h

@@ -9,7 +9,7 @@
  *                         University of Stuttgart.  All rights reserved.
  * Copyright (c) 2004-2005 The Regents of the University of California.
  *                         All rights reserved.
- * Copyright (c) 2006      Sun Microsystems, Inc.  All rights reserved.
+ * Copyright (c) 2006-2007 Sun Microsystems, Inc.  All rights reserved.
  *                         Use is subject to license terms.
  * $COPYRIGHT$
  * 
@@ -129,10 +129,10 @@ int orte_pls_gridengine_signal_proc(const orte_process_name_t*, int32_t);
  */
 struct orte_pls_gridengine_component_t {
     orte_pls_base_component_t super;
-    orte_jobid_t jobid;
+    bool debug;
+    bool verbose;
+    bool daemonize_orted;
     int priority;
-    int verbose;
-    int debug;
     char* orted;
 };
 typedef struct orte_pls_gridengine_component_t orte_pls_gridengine_component_t;

orte/mca/pls/gridengine/pls_gridengine_component.c

@@ -9,7 +9,7 @@
  *                         University of Stuttgart.  All rights reserved.
  * Copyright (c) 2004-2005 The Regents of the University of California.
  *                         All rights reserved.
- * Copyright (c) 2006      Sun Microsystems, Inc.  All rights reserved.
+ * Copyright (c) 2006-2007 Sun Microsystems, Inc.  All rights reserved.
  *                         Use is subject to license terms.
  * $COPYRIGHT$
  * 
@@ -103,21 +103,28 @@ orte_pls_gridengine_component_open - open component and register all parameters
 */
 int orte_pls_gridengine_component_open(void)
 {
+    int tmp;
     mca_base_component_t *c = &mca_pls_gridengine_component.super.pls_version;
 
     mca_base_param_reg_int(c, "debug",
         "Enable debugging of gridengine pls component",
-        false, false, 0, &mca_pls_gridengine_component.debug);
+        false, false, false, &tmp);
+    mca_pls_gridengine_component.debug = OPAL_INT_TO_BOOL(tmp);
     mca_base_param_reg_int(c, "verbose",
         "Enable verbose output of the gridengine qrsh -inherit command",
-        false, false, 0, &mca_pls_gridengine_component.verbose);
+        false, false, false, &tmp);
+    mca_pls_gridengine_component.verbose = OPAL_INT_TO_BOOL(tmp);
     mca_base_param_reg_int(c, "priority",
         "Priority of the gridengine pls component",
         false , false, 100, &mca_pls_gridengine_component.priority);
     mca_base_param_reg_string(c, "orted",
         "The command name that the gridengine pls component will invoke for the ORTE daemon",
         false, false, "orted", &mca_pls_gridengine_component.orted);
-
+    mca_base_param_reg_int(c, "daemonize_orted",
+        "Daemonize orted on remote nodes",
+        false, false, false, &tmp);
+    mca_pls_gridengine_component.daemonize_orted = OPAL_INT_TO_BOOL(tmp);
+    
     return ORTE_SUCCESS;
 }
 

orte/mca/pls/gridengine/pls_gridengine_module.c

@@ -255,13 +255,20 @@ int orte_pls_gridengine_launch_job(orte_jobid_t jobid)
 
     node_name_index1 = argc;
     opal_argv_append(&argc, &argv, "<template>");
-
-    /* add the orted daemon in command and
-     * force orted in the same ptree as sge_shephard with no daemonize */
+    
+    /* add the daemon command */
     orted_index = argc;
     opal_argv_append(&argc, &argv, mca_pls_gridengine_component.orted);
-    opal_argv_append(&argc, &argv, "--no-daemonize");
 
+    /* By default, --no-daemonize will be used and orted will be forced to
+     * to stay in the same ptree as sge_shephard. The problem with
+     * --no-daemonize is that the qrsh -inherit connections will stay
+     * persistent for the whole duration of the task to the remote nodes,
+     * which may not be ideal for large number of nodes */
+    if (! mca_pls_gridengine_component.daemonize_orted) {
+        opal_argv_append(&argc, &argv, "--no-daemonize");
+    }
+    
     /* Add basic orted command line options */
     orte_pls_base_orted_append_basic_args(&argc, &argv,
                                           &proc_name_index,
@@ -542,9 +549,9 @@ int orte_pls_gridengine_launch_job(orte_jobid_t jobid)
             }
             
             /* indicate this daemon has been launched in case anyone is sitting on that trigger */ 
-            if (ORTE_SUCCESS != (rc = orte_smr.set_proc_state(rmaps_node->daemon, ORTE_PROC_STATE_LAUNCHED, 0))) { 
-                ORTE_ERROR_LOG(rc); 
- 	        goto cleanup; 
+            if (ORTE_SUCCESS != (rc = orte_smr.set_proc_state(rmaps_node->daemon, ORTE_PROC_STATE_LAUNCHED, 0))) {
+                ORTE_ERROR_LOG(rc);
+                goto cleanup;
             }
             
             /* setup callback on sigchild - wait until setup above is complete