File pacemaker-crmd-check-too-many-stonith-failures.patch of Package pacemaker.14737

commit 100dd5fda476ef526ac1964260252b30864d5ca7
Author: Ken Gaillot <kgaillot@redhat.com>
Date:   Fri Apr 7 16:51:29 2017 -0500

    Fix: crmd: check for too many stonith failures only when aborting for that reason
    
    Previously, crmd would check for too many stonith failures whenever aborting
    a transition. This would lead to a new transition not being triggered when
    aborting for some other unrelated reason, such as a configuration change.
    
    Now, crmd checks for too many stonith failures only when aborting due to a new
    stonith failure.

diff --git a/crmd/crmd_utils.h b/crmd/crmd_utils.h
index a1aaad32d..d2f8eb2ab 100644
--- a/crmd/crmd_utils.h
+++ b/crmd/crmd_utils.h
@@ -100,8 +100,8 @@ int crmd_join_phase_count(enum crm_join_phase phase);
 void crmd_join_phase_log(int level);
 
 const char *get_timer_desc(fsa_timer_t * timer);
-gboolean too_many_st_failures(void);
 void st_fail_count_reset(const char * target);
+void abort_for_stonith_failure(xmlNode *reason);
 void crmd_peer_down(crm_node_t *peer, bool full);
 
 /* Convenience macro for registering a CIB callback
diff --git a/crmd/te_actions.c b/crmd/te_actions.c
index a8ad86f94..66dd16ebc 100644
--- a/crmd/te_actions.c
+++ b/crmd/te_actions.c
@@ -726,15 +726,11 @@ notify_crmd(crm_graph_t * graph)
         case tg_restart:
             type = "restart";
             if (fsa_state == S_TRANSITION_ENGINE) {
-                if (too_many_st_failures() == FALSE) {
-                    if (transition_timer->period_ms > 0) {
-                        crm_timer_stop(transition_timer);
-                        crm_timer_start(transition_timer);
-                    } else {
-                        event = I_PE_CALC;
-                    }
+                if (transition_timer->period_ms > 0) {
+                    crm_timer_stop(transition_timer);
+                    crm_timer_start(transition_timer);
                 } else {
-                    event = I_TE_SUCCESS;
+                    event = I_PE_CALC;
                 }
 
             } else if (fsa_state == S_POLICY_ENGINE) {
diff --git a/crmd/te_callbacks.c b/crmd/te_callbacks.c
index 6c0670c96..a0aa0813f 100644
--- a/crmd/te_callbacks.c
+++ b/crmd/te_callbacks.c
@@ -635,7 +635,7 @@ struct st_fail_rec {
     int count;
 };
 
-gboolean
+static gboolean
 too_many_st_failures(void)
 {
     GHashTableIter iter;
@@ -694,6 +694,26 @@ st_fail_count_increment(const char *target, int rc)
     }
 }
 
+/*!
+ * \internal
+ * \brief Abort transition due to stonith failure
+ *
+ * \param[in] reason  Failed stonith action XML, or NULL
+ */
+void
+abort_for_stonith_failure(xmlNode *reason)
+{
+    enum transition_action abort_action = tg_restart;
+
+    /* If stonith repeatedly fails, we eventually give up on starting a new
+     * transition for that reason.
+     */
+    if (too_many_st_failures()) {
+        abort_action = tg_stop;
+    }
+    abort_transition(INFINITY, abort_action, "Stonith failed", reason);
+}
+
 void
 tengine_stonith_callback(stonith_t * stonith, stonith_callback_data_t * data)
 {
@@ -759,7 +779,7 @@ tengine_stonith_callback(stonith_t * stonith, stonith_callback_data_t * data)
         action->failed = TRUE;
         crm_notice("Stonith operation %d for %s failed (%s): aborting transition.",
                    call_id, target, pcmk_strerror(rc));
-        abort_transition(INFINITY, tg_restart, "Stonith failed", NULL);
+        abort_for_stonith_failure(NULL);
         st_fail_count_increment(target, rc);
     }
 
diff --git a/crmd/te_utils.c b/crmd/te_utils.c
index 3b67afe15..4603307bd 100644
--- a/crmd/te_utils.c
+++ b/crmd/te_utils.c
@@ -162,7 +162,7 @@ fail_incompletable_stonith(crm_graph_t * graph)
 
     if (last_action != NULL) {
         crm_warn("STONITHd failure resulted in un-runnable actions");
-        abort_transition(INFINITY, tg_restart, "Stonith failure", last_action);
+        abort_for_stonith_failure(last_action);
         return TRUE;
     }