|
Galera cluster master-master async replication goes into hang mode when executing following test case.
– galera master 1
CREATE TABLE t1 (f1 INTEGER PRIMARY KEY AUTO_INCREMENT, f2 INTEGER);
insert into t1(f1,f2) select seq,seq from seq_1_to_10000000;
– galera master 2
ALTER TABLE t1 DROP COLUMN f2;
– galera master 1 (run this before finishing ALTER on master 2)
delete from t1 limit 10;
master 2 goes into hang mode, DDL/DMLs are not proceeding further on master 2
MariaDB [(none)]> show processlist;
|
+----+-------------+-----------------+------+-------------+------+---------------------------------------------------------------+-------------------------------+----------+
|
| Id | User | Host | db | Command | Time | State | Info | Progress |
|
+----+-------------+-----------------+------+-------------+------+---------------------------------------------------------------+-------------------------------+----------+
|
| 1 | system user | | NULL | Sleep | 302 | wsrep aborter idle | NULL | 0.000 |
|
| 2 | system user | | NULL | Sleep | 302 | closing tables | NULL | 0.000 |
|
| 7 | system user | | NULL | Sleep | 302 | wsrep applier idle | NULL | 0.000 |
|
| 10 | root | localhost | test | Query | 125 | Waiting for TOI DDL | ALTER TABLE t1 DROP COLUMN f2 | 0.000 |
|
| 11 | system user | | NULL | Slave_IO | 264 | Waiting for master to send event | NULL | 0.000 |
|
| 13 | repl | localhost:56246 | NULL | Binlog Dump | 255 | Master has sent all binlog to slave; waiting for more updates | NULL | 0.000 |
|
| 14 | root | localhost | test | Field List | 9 | Waiting for table metadata lock | NULL | 0.000 |
|
| 15 | root | localhost | NULL | Query | 0 | starting | show processlist | 0.000 |
|
+----+-------------+-----------------+------+-------------+------+---------------------------------------------------------------+-------------------------------+----------+
|
8 rows in set (0.000 sec)
|
|
MariaDB [(none)]>
|
|