MDEV-22543 : Galera SST donation fails, FLUSH TABLES WITH READ LOCK t…

…imes out

During SST we need to let FTWRL to use normal timeout method
even when client is disconnected.

mysql-test/suite/galera/r/mdev-22543.result

@@ -0,0 +1,17 @@
+connection node_1;
+connection node_2;
+connection node_1;
+CREATE TABLE t1 (f1 INT PRIMARY KEY, f2 INT);
+INSERT INTO t1 VALUES (1, 1);
+SET DEBUG_SYNC = "before_lock_tables_takes_lock SIGNAL sync_point_reached WAIT_FOR sync_point_continue";
+UPDATE t1 SET f2 = 2 WHERE f1 = 1;
+connection node_1_ctrl;
+SET DEBUG_SYNC = "now WAIT_FOR sync_point_reached";
+connection node_2;
+connection node_1_ctrl;
+SET DEBUG_SYNC = "now SIGNAL sync_point_continue";
+connection node_1;
+SET DEBUG_SYNC = "RESET";
+connection node_2;
+connection node_1;
+DROP TABLE t1;

mysql-test/suite/galera/t/mdev-22543.test

@@ -0,0 +1,58 @@
+# The test verifies that the FLUSH TABLES WITH READ LOCK does not
+# time out if it needs to wait for another MDL lock for short duration
+# during SST donation.
+
+--source include/galera_cluster.inc
+--source include/have_debug.inc
+--source include/have_debug_sync.inc
+
+--let $node_1 = node_1
+--let $node_2 = node_2
+--source include/auto_increment_offset_save.inc
+
+--let $galera_connection_name = node_1_ctrl
+--let $galera_server_number = 1
+--source include/galera_connect.inc
+
+#
+# Run UPDATE on node_1 and make it block before table locks are taken.
+# This should block FTWRL.
+#
+--connection node_1
+CREATE TABLE t1 (f1 INT PRIMARY KEY, f2 INT);
+INSERT INTO t1 VALUES (1, 1);
+SET DEBUG_SYNC = "before_lock_tables_takes_lock SIGNAL sync_point_reached WAIT_FOR sync_point_continue";
+--send UPDATE t1 SET f2 = 2 WHERE f1 = 1
+
+--connection node_1_ctrl
+SET DEBUG_SYNC = "now WAIT_FOR sync_point_reached";
+
+#
+# Restart node_2, force SST.
+#
+--connection node_2
+--source include/shutdown_mysqld.inc
+--remove_file $MYSQLTEST_VARDIR/mysqld.2/data/grastate.dat
+# Restart without waiting. The UPDATE should block FTWRL on node_1,
+# so the SST cannot be completed and node_2 cannot join before
+# UPDATE connection is signalled to continue.
+--exec echo "restart:$start_mysqld_params" > $_expect_file_name
+# If the bug is present, FTWRL times out on node_1 in couple of
+# seconds and node_2 fails to join.
+--sleep 10
+
+--connection node_1_ctrl
+SET DEBUG_SYNC = "now SIGNAL sync_point_continue";
+
+--connection node_1
+--reap
+SET DEBUG_SYNC = "RESET";
+
+--connection node_2
+--enable_reconnect
+--source include/wait_until_connected_again.inc
+
+--connection node_1
+DROP TABLE t1;
+
+--source include/auto_increment_offset_restore.inc

sql/mdl.cc

@@ -25,6 +25,7 @@
 #include <mysql/psi/mysql_stage.h>
 #include "wsrep_mysqld.h"
 #include "wsrep_thd.h"
+#include "wsrep_sst.h"
 
 #ifdef HAVE_PSI_INTERFACE
 static PSI_mutex_key key_MDL_wait_LOCK_wait_status;
@@ -2137,18 +2138,26 @@ MDL_context::acquire_lock(MDL_request *mdl_request, double lock_wait_timeout)
     wait_status= m_wait.timed_wait(m_owner, &abs_shortwait, FALSE,
                                    mdl_request->key.get_wait_state_name());
 
+    THD* thd= m_owner->get_thd();
+
     if (wait_status != MDL_wait::EMPTY)
       break;
     /* Check if the client is gone while we were waiting. */
-    if (! thd_is_connected(m_owner->get_thd()))
+    if (! thd_is_connected(thd))
     {
-      /*
-       * The client is disconnected. Don't wait forever:
-       * assume it's the same as a wait timeout, this
-       * ensures all error handling is correct.
-       */
-      wait_status= MDL_wait::TIMEOUT;
-      break;
+#if defined(WITH_WSREP) && !defined(EMBEDDED_LIBRARY)
+      // During SST client might not be connected
+      if (!wsrep_is_sst_progress())
+#endif
+      {
+        /*
+         * The client is disconnected. Don't wait forever:
+         * assume it's the same as a wait timeout, this
+         * ensures all error handling is correct.
+         */
+        wait_status= MDL_wait::TIMEOUT;
+        break;
+      }
     }
 
     mysql_prlock_wrlock(&lock->m_rwlock);

sql/wsrep_sst.cc

@@ -192,6 +192,7 @@ bool wsrep_before_SE()
 
 static bool            sst_complete = false;
 static bool            sst_needed   = false;
+static bool            sst_in_progress = false;
 
 #define WSREP_EXTEND_TIMEOUT_INTERVAL 30
 #define WSREP_TIMEDWAIT_SECONDS 10
@@ -1542,7 +1543,10 @@ static void* sst_donor_thread (void* a)
   char         out_buf[out_len];
 
   wsrep_uuid_t  ret_uuid= WSREP_UUID_UNDEFINED;
-  wsrep_seqno_t ret_seqno= WSREP_SEQNO_UNDEFINED; // seqno of complete SST
+  // seqno of complete SST
+  wsrep_seqno_t ret_seqno= WSREP_SEQNO_UNDEFINED;
+  // SST is now in progress
+  sst_in_progress= true;
 
   wsp::thd thd(FALSE); // we turn off wsrep_on for this THD so that it can
                        // operate with wsrep_ready == OFF
@@ -1644,6 +1648,8 @@ static void* sst_donor_thread (void* a)
   wsrep->sst_sent (wsrep, &state_id, -err);
   proc.wait();
 
+  sst_in_progress= false;
+
   return NULL;
 }
 
@@ -1818,3 +1824,8 @@ void wsrep_SE_initialized()
 {
   SE_initialized = true;
 }
+
+bool wsrep_is_sst_progress()
+{
+  return (sst_in_progress);
+}

sql/wsrep_sst.h

@@ -74,12 +74,14 @@ extern void wsrep_SE_init_grab();   /*! grab init critical section */
 extern void wsrep_SE_init_wait();   /*! wait for SE init to complete */
 extern void wsrep_SE_init_done();   /*! signal that SE init is complte */
 extern void wsrep_SE_initialized(); /*! mark SE initialization complete */
+extern bool wsrep_is_sst_progress();
 
 #else
 #define wsrep_SE_initialized() do { } while(0)
 #define wsrep_SE_init_grab() do { } while(0)
 #define wsrep_SE_init_done() do { } while(0)
 #define wsrep_sst_continue() (0)
+#define wsrep_is_sst_progress() (0)
 
 #endif /* WITH_WSREP */
 #endif /* WSREP_SST_H */