Merge pull request open-mpi#608 from rhc54/cmr2.0/rsh

Now that we have an "isolated" PLM component, we cannot just let rsh …

Author: jsquyres

opal/mca/base/mca_base_components_select.c

@@ -4,6 +4,7 @@
  *                         Corporation.  All rights reserved.
  * Copyright (c) 2015      Los Alamos National Security, LLC. All rights
  *                         reserved.
+ * Copyright (c) 2015      Intel, Inc. All rights reserved.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -38,6 +39,7 @@ int mca_base_select(const char *type_name, int output_id,
     mca_base_component_t *component = NULL;
     mca_base_module_t *module = NULL;
     int priority = 0, best_priority = INT32_MIN;
+    int rc;
 
     *best_module = NULL;
     *best_component = NULL;
@@ -70,7 +72,18 @@ int mca_base_select(const char *type_name, int output_id,
                              "mca:base:select:(%5s) Querying component [%s]",
                              type_name, component->mca_component_name);
 
-        component->mca_query_component(&module, &priority);
+        rc = component->mca_query_component(&module, &priority);
+        if (OPAL_ERR_FATAL == rc) {
+            /* a fatal error was detected by this component - e.g., the
+             * user specified a required element and the component could
+             * not find it. In this case, we must not continue as we might
+             * find some other component that could run, causing us to do
+             * something the user didn't want */
+             return rc;
+        } else if (OPAL_SUCCESS != rc) {
+            /* silently skip this component */
+            continue;
+        }
 
         /*
          * If no module was returned, then skip component

orte/mca/ess/hnp/ess_hnp_module.c

@@ -222,31 +222,26 @@ static int rte_init(void)
      * process stats if requested
      */
     if (ORTE_SUCCESS != (ret = mca_base_framework_open(&opal_pstat_base_framework, 0))) {
-        ORTE_ERROR_LOG(ret);
         error = "opal_pstat_base_open";
         goto error;
     }
     if (ORTE_SUCCESS != (ret = opal_pstat_base_select())) {
-        ORTE_ERROR_LOG(ret);
         error = "opal_pstat_base_select";
         goto error;
     }
 
     /* open and setup the state machine */
     if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_state_base_framework, 0))) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_state_base_open";
         goto error;
     }
     if (ORTE_SUCCESS != (ret = orte_state_base_select())) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_state_base_select";
         goto error;
     }
 
     /* open the errmgr */
     if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_errmgr_base_framework, 0))) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_errmgr_base_open";
         goto error;
     }
@@ -257,27 +252,27 @@ static int rte_init(void)
      * first and select that component.
      */
     if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_plm_base_framework, 0))) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_plm_base_open";
         goto error;
     }
 
     if (ORTE_SUCCESS != (ret = orte_plm_base_select())) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_plm_base_select";
+        if (ORTE_ERR_FATAL == ret) {
+            /* we already output a show_help - so keep down the verbage */
+            ret = ORTE_ERR_SILENT;
+        }
         goto error;
     }
     /* if we were spawned by a singleton, our jobid was given to us */
     if (NULL != orte_ess_base_jobid) {
         if (ORTE_SUCCESS != (ret = orte_util_convert_string_to_jobid(&ORTE_PROC_MY_NAME->jobid, orte_ess_base_jobid))) {
-            ORTE_ERROR_LOG(ret);
             error = "convert_string_to_jobid";
             goto error;
         }
         ORTE_PROC_MY_NAME->vpid = 0;
     } else {
         if (ORTE_SUCCESS != (ret = orte_plm.set_hnp_name())) {
-            ORTE_ERROR_LOG(ret);
             error = "orte_plm_set_hnp_name";
             goto error;
         }
@@ -304,7 +299,6 @@ static int rte_init(void)
                                                     orte_process_info.tmpdir_base,
                                                     orte_process_info.nodename, NULL,
                                                     ORTE_PROC_MY_NAME))) {
-            ORTE_ERROR_LOG(ret);
             error = "orte_session_dir define";
             goto error;
         }
@@ -318,7 +312,6 @@ static int rte_init(void)
                                                     orte_process_info.tmpdir_base,
                                                     orte_process_info.nodename, NULL,
                                                     ORTE_PROC_MY_NAME))) {
-            ORTE_ERROR_LOG(ret);
             error = "orte_session_dir";
             goto error;
         }
@@ -330,12 +323,10 @@ static int rte_init(void)
      * OOB Layer
      */
     if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_oob_base_framework, 0))) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_oob_base_open";
         goto error;
     }
     if (ORTE_SUCCESS != (ret = orte_oob_base_select())) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_oob_base_select";
         goto error;
     }
@@ -344,18 +335,15 @@ static int rte_init(void)
      * Runtime Messaging Layer
      */
     if (ORTE_SUCCESS != (ret = mca_base_framework_open(&orte_rml_base_framework, 0))) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_rml_base_open";
         goto error;
     }
     if (ORTE_SUCCESS != (ret = orte_rml_base_select())) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_rml_base_select";
         goto error;
     }
 
     if (ORTE_SUCCESS != (ret = orte_errmgr_base_select())) {
-        ORTE_ERROR_LOG(ret);
         error = "orte_errmgr_base_select";
         goto error;
     }

orte/mca/plm/base/plm_base_select.c

@@ -11,6 +11,7 @@
  *                         All rights reserved.
  * Copyright (c) 2011-2013 Los Alamos National Security, LLC.
  *                         All rights reserved.
+* Copyright (c) 2015      Intel, Inc. All rights reserved.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -32,43 +33,26 @@
 
 
 /**
-* Function for selecting one component from all those that are
+ * Function for selecting one component from all those that are
  * available.
  */
 
 int orte_plm_base_select(void)
 {
-    int exit_status = ORTE_SUCCESS;
+    int rc;
     orte_plm_base_component_t *best_component = NULL;
     orte_plm_base_module_t *best_module = NULL;
 
     /*
      * Select the best component
      */
-    if( OPAL_SUCCESS != mca_base_select("plm", orte_plm_base_framework.framework_output,
-                                        &orte_plm_base_framework.framework_components,
-                                        (mca_base_module_t **) &best_module,
-                                        (mca_base_component_t **) &best_component) ) {
-        /* This will only happen if no component was selected
-         *
-         * If we didn't find one, and we are a daemon, then default to retaining the proxy.
-         * Otherwise, if we didn't find one to select, that is unacceptable.
-         */
-        if (ORTE_PROC_IS_DAEMON) {
-            /* don't record a selected component or flag selected
-             * so we finalize correctly - just leave the plm alone
-             * as it defaults to pointing at the proxy
-             */
-            goto cleanup;
-        } else {
-            exit_status = ORTE_ERR_NOT_FOUND;
-            goto cleanup;
-        }
+    if (OPAL_SUCCESS == (rc = mca_base_select("plm", orte_plm_base_framework.framework_output,
+                                              &orte_plm_base_framework.framework_components,
+                                              (mca_base_module_t **) &best_module,
+                                              (mca_base_component_t **) &best_component))) {
+        /* Save the winner */
+        orte_plm = *best_module;
     }
 
-    /* Save the winner */
-    orte_plm = *best_module;
-
- cleanup:
-    return exit_status;
+    return rc;
 }

orte/mca/plm/rsh/help-plm-rsh.txt

@@ -10,6 +10,7 @@
 #                         University of Stuttgart.  All rights reserved.
 # Copyright (c) 2004-2005 The Regents of the University of California.
 #                         All rights reserved.
+# Copyright (c) 2015      Intel, Inc. All rights reserved.
 # $COPYRIGHT$
 #
 # Additional copyrights may follow
@@ -18,6 +19,14 @@
 #
 # This is the US/English general help file for Open RTE's orterun.
 #
+[agent-not-found]
+The value of the MCA parameter "plm_rsh_agent" was set to a path
+that could not be found:
+
+  plm_rsh_agent: %s
+
+Please either unset the parameter, or check that the path is correct
+#
 [no-local-orted]
 The rsh PLS component was not able to find the executable "orted" in
 your PATH or in the directory where Open MPI/OpenRTE was initially installed,

orte/mca/plm/rsh/plm_rsh.h

@@ -59,6 +59,8 @@ struct orte_plm_rsh_component_t {
     bool no_tree_spawn;
     int num_concurrent;
     char *agent;
+    char *agent_path;
+    char **agent_argv;
     bool assume_same_shell;
     bool pass_environ_mca_params;
     char *ssh_args;