MDEV-21153 Replica nodes crash due to indexed virtual columns and FK cascading deletebb-10.3-MDEV-21153

Fix for MDEV-23033 fixes a problem in replication applying of transactions, which contain cascading foreign key delete for a table, which has indexed virtual column.
This fix adds slave_fk_event_map flag for table, to mark when the prelocking is needed for applying of a transaction.
See commit 608b0ee52ef3e854ce14a407e64e936adbbeba23 for more details.
However, this fix is targeted for async replication only, Rows_log_event::do_apply_event() has condition to rule out galera replication from the fix domain, and use cases suffering from MDEV-23033 and related MDEV-21153 will fail in galera cluster.

The fix in this commit removes the condition to rule out the setting of slave_fk_event_map flag from galera replication, and makes the fix in MDEV-23033 effective for galera replication as well.

Finally, a mtr test for virtual column support has been added. galera.galera_virtual_column.test has as first test a scenario from MDEV-21153

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

3 files changed, 60 insertions, 1 deletions

diff --git a/mysql-test/suite/galera/r/galera_virtual_column.result b/mysql-test/suite/galera/r/galera_virtual_column.result
new file mode 100644
index 00000000000..c4b425964c5
--- /dev/null
+++ b/mysql-test/suite/galera/r/galera_virtual_column.result
@@ -0,0 +1,17 @@
+connection node_1;
+CREATE TABLE p (id INT UNSIGNED PRIMARY KEY AUTO_INCREMENT) ENGINE = InnoDB;
+CREATE TABLE c (id INT UNSIGNED PRIMARY KEY AUTO_INCREMENT, pid INT UNSIGNED, bitmap TINYINT UNSIGNED NOT NULL DEFAULT 0, bitmap5 TINYINT UNSIGNED GENERATED ALWAYS AS (bitmap&(1<<5)) VIRTUAL, FOREIGN KEY (pid) REFERENCES p (id) ON DELETE CASCADE ON UPDATE CASCADE);
+CREATE INDEX bitmap5 ON c(bitmap5) USING BTREE;
+INSERT INTO p VALUES(1);
+INSERT INTO c(pid) VALUES(1);
+connection node_2;
+connection node_1;
+DELETE FROM p WHERE id=1;
+SELECT * FROM p;
+id
+SELECT * FROM c;
+id	pid	bitmap	bitmap5
+connection node_2;
+connection node_1;
+DROP TABLE c;
+DROP TABLE p;
diff --git a/mysql-test/suite/galera/t/galera_virtual_column.test b/mysql-test/suite/galera/t/galera_virtual_column.test
new file mode 100644
index 00000000000..84e1da024f1
--- /dev/null
+++ b/mysql-test/suite/galera/t/galera_virtual_column.test
@@ -0,0 +1,42 @@
+#
+# This test is for testing virtual columnm support in galera cluster
+#
+--source include/galera_cluster.inc
+--source include/have_innodb.inc
+
+#
+# test case for verifying that cascaded delete in a table with virtual column does not crash slave node
+#
+
+--connection node_1
+
+CREATE TABLE p (id INT UNSIGNED PRIMARY KEY AUTO_INCREMENT) ENGINE = InnoDB;
+CREATE TABLE c (id INT UNSIGNED PRIMARY KEY AUTO_INCREMENT, pid INT UNSIGNED, bitmap TINYINT UNSIGNED NOT NULL DEFAULT 0, bitmap5 TINYINT UNSIGNED GENERATED ALWAYS AS (bitmap&(1<<5)) VIRTUAL, FOREIGN KEY (pid) REFERENCES p (id) ON DELETE CASCADE ON UPDATE CASCADE);
+
+# not sure of this index is needed for the test
+CREATE INDEX bitmap5 ON c(bitmap5) USING BTREE;
+
+INSERT INTO p VALUES(1);
+INSERT INTO c(pid) VALUES(1);
+
+
+--connection node_2
+# wait until both INSERTS have arrived in node_2
+--let $wait_condition = SELECT COUNT(*) = 1 FROM c
+--source include/wait_condition.inc
+
+--connection node_1
+# delete from parent table, it will cascade into child table
+# node_2 might have problem in applying this cascaded delete
+DELETE FROM p WHERE id=1;
+
+SELECT * FROM p;
+SELECT * FROM c;
+
+--connection node_2
+--let $wait_condition = SELECT COUNT(*) = 0 FROM c;
+--source include/wait_condition.inc
+
+--connection node_1
+DROP TABLE c;
+DROP TABLE p;
diff --git a/sql/log_event.cc b/sql/log_event.cc
index 48a25c8ba3d..e344fc8894f 100644
--- a/sql/log_event.cc
+++ b/sql/log_event.cc
@@ -11357,7 +11357,7 @@ int Rows_log_event::do_apply_event(rpl_group_info *rgi)
         tables->trg_event_map= new_trg_event_map;
         lex->query_tables_last= &tables->next_global;
       }
-      else if (!WSREP_ON)
+      else
       {
         tables->slave_fk_event_map= new_trg_event_map;
         lex->query_tables_last= &tables->next_global;