File bsc#1181744-0004-Fix-fence-history-resync-fence-history-after-fenced-.patch of Package pacemaker.21298
From 03c4455fced74f093deb782198b1ba3076e52015 Mon Sep 17 00:00:00 2001
From: Klaus Wenninger <klaus.wenninger@aon.at>
Date: Tue, 18 Jun 2019 14:12:27 +0200
Subject: [PATCH 4/7] Fix: fence-history: resync fence-history after fenced
crash
Setting up a 30s fallback timer to trigger history-sync if the
sync via DC doesn't happen
---
daemons/controld/controld_callbacks.c | 2 +-
daemons/controld/controld_control.c | 2 +
daemons/controld/controld_te_utils.c | 86 +++++++++++++++++++++++----
daemons/controld/controld_transition.h | 3 +-
4 files changed, 79 insertions(+), 14 deletions(-)
Index: pacemaker-2.0.1+20190417.13d370ca9/daemons/controld/controld_callbacks.c
===================================================================
--- pacemaker-2.0.1+20190417.13d370ca9.orig/daemons/controld/controld_callbacks.c
+++ pacemaker-2.0.1+20190417.13d370ca9/daemons/controld/controld_callbacks.c
@@ -210,7 +210,7 @@ peer_update_callback(enum crm_status_typ
} else if(AM_I_DC) {
if (appeared) {
- te_trigger_stonith_history_sync();
+ te_trigger_stonith_history_sync(FALSE);
} else {
erase_status_tag(node->uname, XML_TAG_TRANSIENT_NODEATTRS, cib_scope_local);
}
Index: pacemaker-2.0.1+20190417.13d370ca9/daemons/controld/controld_control.c
===================================================================
--- pacemaker-2.0.1+20190417.13d370ca9.orig/daemons/controld/controld_control.c
+++ pacemaker-2.0.1+20190417.13d370ca9/daemons/controld/controld_control.c
@@ -118,7 +118,12 @@ do_shutdown(long long action,
clear_bit(fsa_input_register, R_ST_REQUIRED);
crm_info("Disconnecting from fencer");
- stonith_api->cmds->disconnect(stonith_api);
+ if (stonith_api->state != stonith_disconnected) {
+ stonith_api->cmds->disconnect(stonith_api);
+ }
+ stonith_api->cmds->remove_notification(stonith_api, T_STONITH_NOTIFY_DISCONNECT);
+ stonith_api->cmds->remove_notification(stonith_api, T_STONITH_NOTIFY_FENCE);
+ stonith_api->cmds->remove_notification(stonith_api, T_STONITH_NOTIFY_HISTORY_SYNCED);
}
}
@@ -272,6 +277,8 @@ crmd_exit(crm_exit_t exit_code)
crm_timer_stop(wait_timer);
crm_timer_stop(recheck_timer);
+ te_cleanup_stonith_history_sync(NULL, TRUE);
+
free(transition_timer); transition_timer = NULL;
free(integration_timer); integration_timer = NULL;
free(finalization_timer); finalization_timer = NULL;
Index: pacemaker-2.0.1+20190417.13d370ca9/daemons/controld/controld_te_utils.c
===================================================================
--- pacemaker-2.0.1+20190417.13d370ca9.orig/daemons/controld/controld_te_utils.c
+++ pacemaker-2.0.1+20190417.13d370ca9/daemons/controld/controld_te_utils.c
@@ -22,7 +22,33 @@
crm_trigger_t *stonith_reconnect = NULL;
static crm_trigger_t *stonith_history_sync_trigger = NULL;
-static mainloop_timer_t *stonith_history_sync_timer = NULL;
+static mainloop_timer_t *stonith_history_sync_timer_short = NULL;
+static mainloop_timer_t *stonith_history_sync_timer_long = NULL;
+
+void
+te_cleanup_stonith_history_sync(stonith_t *st, bool free_timers)
+{
+ if (free_timers) {
+ mainloop_timer_del(stonith_history_sync_timer_short);
+ stonith_history_sync_timer_short = NULL;
+ mainloop_timer_del(stonith_history_sync_timer_long);
+ stonith_history_sync_timer_long = NULL;
+ } else {
+ mainloop_timer_stop(stonith_history_sync_timer_short);
+ mainloop_timer_stop(stonith_history_sync_timer_long);
+ }
+
+ if (st) {
+ st->cmds->remove_notification(st, T_STONITH_NOTIFY_HISTORY_SYNCED);
+ }
+}
+
+static void
+tengine_stonith_history_synced(stonith_t *st, stonith_event_t *st_event)
+{
+ te_cleanup_stonith_history_sync(st, FALSE);
+ crm_debug("Fence-history synced - cancel all timers");
+}
/*
* stonith cleanup list
@@ -163,6 +189,8 @@ fail_incompletable_stonith(crm_graph_t *
static void
tengine_stonith_connection_destroy(stonith_t * st, stonith_event_t * e)
{
+ te_cleanup_stonith_history_sync(st, FALSE);
+
if (is_set(fsa_input_register, R_ST_REQUIRED)) {
crm_crit("Fencing daemon connection failed");
mainloop_set_trigger(stonith_reconnect);
@@ -176,11 +204,12 @@ tengine_stonith_connection_destroy(stoni
/* the client API won't properly reconnect notifications
* if they are still in the table - so remove them
*/
- stonith_api->cmds->remove_notification(st, T_STONITH_NOTIFY_DISCONNECT);
- stonith_api->cmds->remove_notification(st, T_STONITH_NOTIFY_FENCE);
if (stonith_api->state != stonith_disconnected) {
stonith_api->cmds->disconnect(st);
}
+ stonith_api->cmds->remove_notification(stonith_api, T_STONITH_NOTIFY_DISCONNECT);
+ stonith_api->cmds->remove_notification(stonith_api, T_STONITH_NOTIFY_FENCE);
+ stonith_api->cmds->remove_notification(stonith_api, T_STONITH_NOTIFY_HISTORY_SYNCED);
}
if (AM_I_DC) {
@@ -197,6 +226,9 @@ char *te_client_id = NULL;
#endif
static void
+tengine_stonith_history_synced(stonith_t *st, stonith_event_t *st_event);
+
+static void
tengine_stonith_notify(stonith_t * st, stonith_event_t * st_event)
{
if(te_client_id == NULL) {
@@ -345,6 +377,7 @@ do_stonith_history_sync(gpointer user_da
if (stonith_api && (stonith_api->state != stonith_disconnected)) {
stonith_history_t *history = NULL;
+ te_cleanup_stonith_history_sync(stonith_api, FALSE);
stonith_api->cmds->history(stonith_api,
st_opt_sync_call | st_opt_broadcast,
NULL, &history, 5);
@@ -364,11 +397,18 @@ stonith_history_sync_set_trigger(gpointe
}
void
-te_trigger_stonith_history_sync(void)
+te_trigger_stonith_history_sync(bool long_timeout)
{
/* trigger a sync in 5s to give more nodes the
* chance to show up so that we don't create
* unnecessary stonith-history-sync traffic
+ *
+ * the long timeout of 30s is there as a fallback
+ * so that after a successful connection to fenced
+ * we will wait for 30s for the DC to trigger a
+ * history-sync
+ * if this doesn't happen we trigger a sync locally
+ * (e.g. fenced segfaults and is restarted by pacemakerd)
*/
/* as we are finally checking the stonith-connection
@@ -382,14 +422,26 @@ te_trigger_stonith_history_sync(void)
do_stonith_history_sync, NULL);
}
- if(stonith_history_sync_timer == NULL) {
- stonith_history_sync_timer =
- mainloop_timer_add("history_sync", 5000,
- FALSE, stonith_history_sync_set_trigger,
- NULL);
+ if (long_timeout) {
+ if(stonith_history_sync_timer_long == NULL) {
+ stonith_history_sync_timer_long =
+ mainloop_timer_add("history_sync_long", 30000,
+ FALSE, stonith_history_sync_set_trigger,
+ NULL);
+ }
+ crm_info("Fence history will be synchronized cluster-wide within 30 seconds");
+ mainloop_timer_start(stonith_history_sync_timer_long);
+ } else {
+ if(stonith_history_sync_timer_short == NULL) {
+ stonith_history_sync_timer_short =
+ mainloop_timer_add("history_sync_short", 5000,
+ FALSE, stonith_history_sync_set_trigger,
+ NULL);
+ }
+ crm_info("Fence history will be synchronized cluster-wide within 5 seconds");
+ mainloop_timer_start(stonith_history_sync_timer_short);
}
- crm_info("Fence history will be synchronized cluster-wide within 5 seconds");
- mainloop_timer_start(stonith_history_sync_timer);
+
}
gboolean
@@ -437,6 +489,11 @@ te_connect_stonith(gpointer user_data)
stonith_api->cmds->register_notification(stonith_api, T_STONITH_NOTIFY_FENCE,
tengine_stonith_notify);
+ stonith_api->cmds->register_notification(stonith_api,
+ T_STONITH_NOTIFY_HISTORY_SYNCED,
+ tengine_stonith_history_synced);
+
+ te_trigger_stonith_history_sync(TRUE);
crm_trace("Connected");
return TRUE;
Index: pacemaker-2.0.1+20190417.13d370ca9/daemons/controld/controld_transition.h
===================================================================
--- pacemaker-2.0.1+20190417.13d370ca9.orig/daemons/controld/controld_transition.h
+++ pacemaker-2.0.1+20190417.13d370ca9/daemons/controld/controld_transition.h
@@ -70,7 +70,8 @@ extern void abort_transition_graph(int a
extern gboolean te_connect_stonith(gpointer user_data);
-extern void te_trigger_stonith_history_sync(void);
+extern void te_trigger_stonith_history_sync(bool long_timeout);
+extern void te_cleanup_stonith_history_sync(stonith_t *st, bool free_timers);
extern crm_trigger_t *transition_trigger;
extern crm_trigger_t *stonith_reconnect;
