Provide a better help message when we are unable to complete a connection due to a firewall.

cmr=v1.8.3:reviewer=jsquyres

This commit was SVN r32743.

orte/mca/oob/tcp/help-oob-tcp.txt

@@ -64,4 +64,16 @@ value will be ignored.
   Local host: %s
   Value:      %s
   Message:    %s
-  
+#
+[firewall]
+A process or daemon is unable to complete a TCP connection
+to another process:
+
+  Local host:    %s
+  Local proc:    %s
+  Remote host:   %s
+  Remote proc:   %s
+
+This is usually caused by a firewall on the remote host. Please
+check that any firewall (e.g., iptables) has been disabled and
+try again.

orte/mca/oob/tcp/oob_tcp_connection.c

@@ -60,6 +60,7 @@
 #include "opal/mca/event/event.h"
 
 #include "orte/util/name_fns.h"
+#include "orte/util/show_help.h"
 #include "orte/mca/state/state.h"
 #include "orte/runtime/orte_globals.h"
 #include "orte/mca/errmgr/errmgr.h"
@@ -254,12 +255,11 @@ void mca_oob_tcp_peer_try_connect(int fd, short args, void *cbdata)
         /* no address succeeded, so we cannot reach this peer */
         peer->state = MCA_OOB_TCP_FAILED;
         host = orte_get_proc_hostname(&(peer->name));
-        opal_output_verbose(OOB_TCP_DEBUG_CONNECT, orte_oob_base_framework.framework_output,
-                            "%s orte_tcp_peer_try_connect: "
-                            "Connection to proc %s on node %s failed",
-                            ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
-                            ORTE_NAME_PRINT(&peer->name),
-                            (NULL == host) ? "NULL" : host);
+        orte_show_help("help-oob-tcp.txt", "firewall",
+                       orte_process_info.nodename,
+                       ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
+                       (NULL == host) ? "NULL" : host,
+                       ORTE_NAME_PRINT(&peer->name));
         /* let the TCP component know that this module failed to make
          * the connection so it can do some bookkeeping and fail back
          * to the OOB level so another component can try. This will activate

orte/runtime/orte_globals.c

@@ -507,14 +507,15 @@ char* orte_get_proc_hostname(orte_process_name_t *proc)
     opal_list_t myvals;
     opal_value_t *kv;
 
+    /* don't bother error logging any not-found situations
+     * as the layer above us will have something to say
+     * about it */
     if (ORTE_PROC_IS_DAEMON || ORTE_PROC_IS_HNP) {
         /* look it up on our arrays */
         if (NULL == (proct = orte_get_proc_object(proc))) {
-            ORTE_ERROR_LOG(ORTE_ERR_NOT_FOUND);
             return NULL;
         }
         if (NULL == proct->node || NULL == proct->node->name) {
-            ORTE_ERROR_LOG(ORTE_ERR_NOT_FOUND);
             return NULL;
         }
         return proct->node->name;
@@ -526,7 +527,6 @@ char* orte_get_proc_hostname(orte_process_name_t *proc)
                                                 (opal_identifier_t*)proc,
                                                 OPAL_DSTORE_HOSTNAME,
                                                 &myvals))) {
-        ORTE_ERROR_LOG(rc);
         OPAL_LIST_DESTRUCT(&myvals);
         return NULL;
     }