MDEV-23328 Server hang due to Galera lock conflict resolution

Mutex order violation when wsrep bf thread kills a conflicting trx,
the stack is

          wsrep_thd_LOCK()
          wsrep_kill_victim()
          lock_rec_other_has_conflicting()
          lock_clust_rec_read_check_and_lock()
          row_search_mvcc()
          ha_innobase::index_read()
          ha_innobase::rnd_pos()
          handler::ha_rnd_pos()
          handler::rnd_pos_by_record()
          handler::ha_rnd_pos_by_record()
          Rows_log_event::find_row()
          Update_rows_log_event::do_exec_row()
          Rows_log_event::do_apply_event()
          Log_event::apply_event()
          wsrep_apply_events()

and mutexes are taken in the order

          lock_sys->mutex -> victim_trx->mutex -> victim_thread->LOCK_thd_data

When a normal KILL statement is executed, the stack is

          innobase_kill_query()
          kill_handlerton()
          plugin_foreach_with_mask()
          ha_kill_query()
          THD::awake()
          kill_one_thread()

        and mutexes are

          victim_thread->LOCK_thd_data -> lock_sys->mutex -> victim_trx->mutex

This patch is the plan D variant for fixing potetial mutex locking
order exercised by BF aborting and KILL command execution.

In this approach, KILL command is replicated as TOI operation.
This guarantees total isolation for the KILL command execution
in the first node: there is no concurrent replication applying
and no concurrent DDL executing. Therefore there is no risk of
BF aborting to happen in parallel with KILL command execution
either. Potential mutex deadlocks between the different mutex
access paths with KILL command execution and BF aborting cannot
therefore happen.

TOI replication is used, in this approach,  purely as means
to provide isolated KILL command execution in the first node.
KILL command should not (and must not) be applied in secondary
nodes. In this patch, we make this sure by skipping KILL
execution in secondary nodes, in applying phase, where we
bail out if applier thread is trying to execute KILL command.
This is effective, but skipping the applying of KILL command
could happen much earlier as well.

This also fixed unprotected calls to wsrep_thd_abort
that will use wsrep_abort_transaction. This is fixed
by holding THD::LOCK_thd_data while we abort transaction.

Reviewed-by: Jan Lindström <jan.lindstrom@mariadb.com>

8 files changed, 122 insertions, 28 deletions

diff --git a/mysql-test/suite/galera/r/galera_UK_conflict.result b/mysql-test/suite/galera/r/galera_UK_conflict.result
index 76649f1b268..2795a86d6a6 100644
--- a/mysql-test/suite/galera/r/galera_UK_conflict.result
+++ b/mysql-test/suite/galera/r/galera_UK_conflict.result
@@ -68,6 +68,9 @@ f1	f2	f3
 10	10	0
 INSERT INTO t1 VALUES (7,7,7);
 INSERT INTO t1 VALUES (8,8,8);
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+7
 SELECT * FROM t1;
 f1	f2	f3
 1	1	0
@@ -78,6 +81,9 @@ f1	f2	f3
 8	8	8
 10	10	0
 connection node_1;
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+7
 SELECT * FROM t1;
 f1	f2	f3
 1	1	0
@@ -85,5 +91,6 @@ f1	f2	f3
 4	4	2
 5	5	2
 7	7	7
+8	8	8
 10	10	0
 DROP TABLE t1;
diff --git a/mysql-test/suite/galera/r/galera_to_error.result b/mysql-test/suite/galera/r/galera_to_error.result
new file mode 100644
index 00000000000..4f985593466
--- /dev/null
+++ b/mysql-test/suite/galera/r/galera_to_error.result
@@ -0,0 +1,30 @@
+connection node_1;
+connection node_2;
+connection node_1;
+call mtr.add_suppression("WSREP: TO isolation failed for: ");
+CREATE TABLE t2(a int not null auto_increment primary key, b int, key(b)) engine=innodb;
+INSERT INTO t2 values (NULL,1),(NULL,2),(NULL,3),(NULL,4),(NULL,5),(NULL,6);
+connect node_1a, 127.0.0.1, root, , test, $NODE_MYPORT_1;
+connection node_1a;
+BEGIN;
+UPDATE t2 set b = b + 20 where b BETWEEN 2 and 5;;
+connect node_1b, 127.0.0.1, root, , test, $NODE_MYPORT_1;
+connection node_1b;
+connection node_2;
+Killing server ...
+connect node_1c, 127.0.0.1, root, , test, $NODE_MYPORT_1;
+connection node_1c;
+CREATE TABLE t1 (f1 INTEGER NOT NULL PRIMARY KEY) ENGINE=InnoDB;
+ERROR 40001: WSREP replication failed. Check your wsrep connection state and retry the query.
+ERROR 40001: WSREP replication failed. Check your wsrep connection state and retry the query.
+CREATE UNIQUE INDEX b2 ON t2(b);
+ERROR 08S01: WSREP has not yet prepared node for application use
+connection node_1;
+disconnect node_1a;
+disconnect node_1b;
+disconnect node_1c;
+connection node_2;
+connect node_2a, 127.0.0.1, root, , test, $NODE_MYPORT_2;
+DROP TABLE t2;
+disconnect node_2;
+disconnect node_1;
diff --git a/mysql-test/suite/galera/r/galera_toi_ddl_fk_insert.result b/mysql-test/suite/galera/r/galera_toi_ddl_fk_insert.result
index 94752ed7c76..a972394f87c 100644
--- a/mysql-test/suite/galera/r/galera_toi_ddl_fk_insert.result
+++ b/mysql-test/suite/galera/r/galera_toi_ddl_fk_insert.result
@@ -21,22 +21,6 @@ connection node_1a;
 connection node_1b;
 connection node_2;
 connection node_2a;
-connection node_1;
-SET SESSION wsrep_sync_wait=15;
-SELECT COUNT(*) FROM parent;
-COUNT(*)
-20001
-SELECT COUNT(*) FROM child;
-COUNT(*)
-10000
-connection node_2;
-SET SESSION wsrep_sync_wait=15;
-SELECT COUNT(*) FROM parent;
-COUNT(*)
-20001
-SELECT COUNT(*) FROM child;
-COUNT(*)
-10000
 DROP TABLE child;
 DROP TABLE parent;
 DROP TABLE ten;
diff --git a/mysql-test/suite/galera/t/galera_UK_conflict.test b/mysql-test/suite/galera/t/galera_UK_conflict.test
index 57bafbf8ae0..83d0e47dc3d 100644
--- a/mysql-test/suite/galera/t/galera_UK_conflict.test
+++ b/mysql-test/suite/galera/t/galera_UK_conflict.test
@@ -140,9 +140,13 @@ SELECT * FROM t1;
 # original state in node 1
 INSERT INTO t1 VALUES (7,7,7);
 INSERT INTO t1 VALUES (8,8,8);
+SELECT COUNT(*) FROM t1;
 SELECT * FROM t1;
 
 --connection node_1
+--let $wait_condition = SELECT COUNT(*) = 7 FROM t1
+--source include/wait_condition.inc
+SELECT COUNT(*) FROM t1;
 SELECT * FROM t1;
 
 DROP TABLE t1;
diff --git a/mysql-test/suite/galera/t/galera_to_error.test b/mysql-test/suite/galera/t/galera_to_error.test
new file mode 100644
index 00000000000..71bfe64c02d
--- /dev/null
+++ b/mysql-test/suite/galera/t/galera_to_error.test
@@ -0,0 +1,71 @@
+#
+# Confirm that with two nodes, killing one causes the other to stop accepting connections
+#
+
+--source include/galera_cluster.inc
+--source include/have_innodb.inc
+
+# Save original auto_increment_offset values.
+--let $node_1=node_1
+--let $node_2=node_2
+--source include/auto_increment_offset_save.inc
+
+--connection node_1
+call mtr.add_suppression("WSREP: TO isolation failed for: ");
+--let $wsrep_cluster_address_orig = `SELECT @@wsrep_cluster_address`
+
+CREATE TABLE t2(a int not null auto_increment primary key, b int, key(b)) engine=innodb;
+INSERT INTO t2 values (NULL,1),(NULL,2),(NULL,3),(NULL,4),(NULL,5),(NULL,6);
+
+--connect node_1a, 127.0.0.1, root, , test, $NODE_MYPORT_1
+--connection node_1a
+BEGIN;
+--send UPDATE t2 set b = b + 20 where b BETWEEN 2 and 5;
+
+#
+# Take thread id for above query
+#
+--connect node_1b, 127.0.0.1, root, , test, $NODE_MYPORT_1
+--connection node_1b
+--let $k_thread = `SELECT ID FROM INFORMATION_SCHEMA.PROCESSLIST WHERE USER = 'root' AND COMMAND = 'Sleep' LIMIT 1`
+
+--connection node_2
+--source include/kill_galera.inc
+
+--connect node_1c, 127.0.0.1, root, , test, $NODE_MYPORT_1
+--connection node_1c
+--error ER_LOCK_DEADLOCK
+CREATE TABLE t1 (f1 INTEGER NOT NULL PRIMARY KEY) ENGINE=InnoDB;
+
+--disable_query_log
+--error ER_LOCK_DEADLOCK
+--eval KILL QUERY $k_thread;
+--enable_query_log
+
+# Reset the master and restart the slave so that post-test checks can run
+--error 1047
+CREATE UNIQUE INDEX b2 ON t2(b);
+
+--connection node_1
+--disconnect node_1a
+--disconnect node_1b
+--disconnect node_1c
+
+--connection node_2
+--source include/start_mysqld.inc
+--source include/wait_until_connected_again.inc
+
+--let $wait_condition = SELECT VARIABLE_VALUE = 2 FROM INFORMATION_SCHEMA.GLOBAL_STATUS WHERE VARIABLE_NAME = 'wsrep_cluster_size'
+--source include/wait_condition.inc
+
+--connect node_2a, 127.0.0.1, root, , test, $NODE_MYPORT_2
+--source include/wait_until_connected_again.inc
+
+DROP TABLE t2;
+
+# Restore original auto_increment_offset values.
+--let $node_2=node_2a
+--source include/auto_increment_offset_restore.inc
+
+--source include/galera_end.inc
+
diff --git a/mysql-test/suite/galera/t/galera_toi_ddl_fk_insert.test b/mysql-test/suite/galera/t/galera_toi_ddl_fk_insert.test
index fadc94d78ff..3b4b427f551 100644
--- a/mysql-test/suite/galera/t/galera_toi_ddl_fk_insert.test
+++ b/mysql-test/suite/galera/t/galera_toi_ddl_fk_insert.test
@@ -54,15 +54,11 @@ INSERT INTO parent VALUES (1, 0);
 --connection node_2a
 --reap
 
---connection node_1
-SET SESSION wsrep_sync_wait=15;
-SELECT COUNT(*) FROM parent;
-SELECT COUNT(*) FROM child;
-
---connection node_2
-SET SESSION wsrep_sync_wait=15;
-SELECT COUNT(*) FROM parent;
-SELECT COUNT(*) FROM child;
+#
+# ALTER TABLE could bf kill one or more of INSERTs to parent, so
+# the actual number of rows in PARENT depends on whether
+# the INSERT is committed before ALTER TABLE is executed
+#
 
 DROP TABLE child;
 DROP TABLE parent;
diff --git a/mysql-test/suite/galera/t/galera_var_auto_inc_control_off.test b/mysql-test/suite/galera/t/galera_var_auto_inc_control_off.test
index c0bbe5af8cf..241b62dbf8c 100644
--- a/mysql-test/suite/galera/t/galera_var_auto_inc_control_off.test
+++ b/mysql-test/suite/galera/t/galera_var_auto_inc_control_off.test
@@ -94,11 +94,13 @@ SELECT * FROM t1;
 --eval SET GLOBAL wsrep_auto_increment_control = $auto_increment_control_orig
 --eval SET GLOBAL auto_increment_increment = $auto_increment_increment_node1
 --eval SET GLOBAL auto_increment_offset = $auto_increment_offset_node1
+--disconnect node_1a
 
 --connection node_2
 --eval SET GLOBAL wsrep_auto_increment_control = $auto_increment_control_orig
 --eval SET GLOBAL auto_increment_increment = $auto_increment_increment_node2
 --eval SET GLOBAL auto_increment_offset = $auto_increment_offset_node2
+--disconnect node_2a
 
 --enable_query_log
 
diff --git a/mysql-test/suite/wsrep/t/variables.test b/mysql-test/suite/wsrep/t/variables.test
index 0cf13650ce0..875315c0e7c 100644
--- a/mysql-test/suite/wsrep/t/variables.test
+++ b/mysql-test/suite/wsrep/t/variables.test
@@ -66,7 +66,7 @@ call mtr.add_suppression("WSREP: Failed to get provider options");
 
 #evalp SET GLOBAL wsrep_provider= '$WSREP_PROVIDER';
 
---replace_regex /.*libgalera_smm.*/libgalera_smm.so/
+--replace_regex /.*libgalera.*/libgalera_smm.so/
 SELECT @@global.wsrep_provider;
 SELECT @@global.wsrep_slave_threads;
 SELECT @@global.wsrep_cluster_address;
@@ -77,7 +77,7 @@ SHOW STATUS LIKE 'wsrep_thread_count';
 
 #evalp SET GLOBAL wsrep_provider= '$WSREP_PROVIDER';
 
---replace_regex /.*libgalera_smm.*/libgalera_smm.so/
+--replace_regex /.*libgalera.*/libgalera_smm.so/
 SELECT @@global.wsrep_provider;
 SELECT @@global.wsrep_cluster_address;
 SELECT @@global.wsrep_on;
@@ -101,7 +101,7 @@ SELECT VARIABLE_VALUE AS EXPECT_1 FROM INFORMATION_SCHEMA.GLOBAL_STATUS WHERE VA
 SELECT VARIABLE_VALUE AS EXPECT_1 FROM INFORMATION_SCHEMA.GLOBAL_STATUS WHERE VARIABLE_NAME = 'wsrep_rollbacker_thread_count';
 SELECT VARIABLE_VALUE AS EXPECT_2 FROM INFORMATION_SCHEMA.GLOBAL_STATUS WHERE VARIABLE_NAME = 'wsrep_thread_count';
 
---replace_regex /.*libgalera_smm.*/libgalera_smm.so/
+--replace_regex /.*libgalera.*/libgalera_smm.so/
 SELECT @@global.wsrep_provider;
 SELECT @@global.wsrep_cluster_address;
 SELECT @@global.wsrep_on;