Debugger daemons spawned upon attachment to a running job need to be treated just like a regular job - they are not "piggybacking" onto an existing launch, and so the orte daemons need to report them just like a regular job launch in order to release from spawn.

Modify the debugger control flag to include "do not monitor" so orterun will not take debugger daemon termination into account when deciding that all jobs are done.

This commit was SVN r23282.

orte/runtime/orte_globals.h

@@ -292,7 +292,7 @@ typedef uint8_t orte_job_controls_t;
 
 #define ORTE_JOB_CONTROL_LOCAL_SLAVE        0x01
 #define ORTE_JOB_CONTROL_NON_ORTE_JOB       0x02
-#define ORTE_JOB_CONTROL_DEBUGGER_DAEMON    0x04
+#define ORTE_JOB_CONTROL_DEBUGGER_DAEMON    0x14
 #define ORTE_JOB_CONTROL_FORWARD_OUTPUT     0x08
 #define ORTE_JOB_CONTROL_DO_NOT_MONITOR     0x10
 #define ORTE_JOB_CONTROL_FORWARD_COMM       0x20

orte/tools/orterun/debuggers.c

@@ -481,6 +481,7 @@ static void check_debugger(int fd, short event, void *arg)
     orte_app_context_t *app;
     int rc;
     int32_t ljob;
+    orte_job_t *jdata;
 
     if (MPIR_being_debugged || orte_debugger_test_attach) {
         if (orte_debug_flag) {
@@ -504,21 +505,21 @@ static void check_debugger(int fd, short event, void *arg)
              * so we do not use the global orte_debugger_daemon
              * as this is reserved for co-location upon startup
              */
-            orte_debugger_daemon = OBJ_NEW(orte_job_t);
+            jdata = OBJ_NEW(orte_job_t);
             /* create a jobid for these daemons - this is done solely
              * to avoid confusing the rest of the system's bookkeeping
              */
-            orte_plm_base_create_jobid(orte_debugger_daemon);
+            orte_plm_base_create_jobid(jdata);
             /* flag the job as being debugger daemons */
-            orte_debugger_daemon->controls |= ORTE_JOB_CONTROL_DEBUGGER_DAEMON;
+            jdata->controls |= ORTE_JOB_CONTROL_DEBUGGER_DAEMON;
             /* unless directed, we do not forward output */
             if (!MPIR_forward_output) {
-                orte_debugger_daemon->controls &= ~ORTE_JOB_CONTROL_FORWARD_OUTPUT;
+                jdata->controls &= ~ORTE_JOB_CONTROL_FORWARD_OUTPUT;
             }
-            orte_debugger_daemon->num_procs = orte_process_info.num_procs;
+            jdata->num_procs = orte_process_info.num_procs;
             /* add it to the global job pool */
-            ljob = ORTE_LOCAL_JOBID(orte_debugger_daemon->jobid);
-            opal_pointer_array_set_item(orte_job_data, ljob, orte_debugger_daemon);
+            ljob = ORTE_LOCAL_JOBID(jdata->jobid);
+            opal_pointer_array_set_item(orte_job_data, ljob, jdata);
             /* create an app_context for the debugger daemon */
             app = OBJ_NEW(orte_app_context_t);
             if (NULL != orte_debugger_test_daemon) {
@@ -528,10 +529,15 @@ static void check_debugger(int fd, short event, void *arg)
             }
             opal_argv_append_nosize(&app->argv, app->app);
             build_debugger_args(app);
-            opal_pointer_array_add(orte_debugger_daemon->apps, &app->super);
-            orte_debugger_daemon->num_apps = 1;
+            opal_pointer_array_add(jdata->apps, &app->super);
+            jdata->num_apps = 1;
+            /* setup the mapping policy to bynode so we get one
+             * daemon on each node
+             */
+            jdata->map = OBJ_NEW(orte_job_map_t);
+            jdata->map->policy = ORTE_MAPPING_BYNODE;
             /* now go ahead and spawn this job */
-            if (ORTE_SUCCESS != (rc = orte_plm.spawn(NULL))) {
+            if (ORTE_SUCCESS != (rc = orte_plm.spawn(jdata))) {
                 ORTE_ERROR_LOG(rc);
             }
         }