Xtrabackup appears' flush no '_ WRITE_ TO_ BINLOG TABLES‘: 1205 (HY000) Lock wait timeout exceeded；

xtrabackup Backup MySQL when , You may fail to get the lock and timeout , Here is the solution .

The phenomenon

xtrabackup Backup MySQL The database has a timeout error ：

Error: failed to execute query 'FLUSH NO_WRITE_TO_BINLOG TABLES': 1205 (HY000) Lock wait timeout exceeded; try restarting transaction
...
Error: failed to execute query 'FLUSH NO_WRITE_TO_BINLOG TABLES': 1205 (HY000) Lock wait timeout exceeded; try restarting transaction

resolvent

There are two solutions , One is modification xtrabackup Parameters of , The other is to kill threads that execute for a long time .

Kill long SQL The thread of ：

Query threads with long execution time ：

mysql> select * from processlist where command not like 'Sleep' order by time desc limit 3;
+-----+-----------------+-----------+--------+---------+--------+-------------------------+-----------------------------------------------+
| ID  | USER            | HOST      | DB     | COMMAND | TIME   | STATE                   | INFO                                          |
+-----+-----------------+-----------+--------+---------+--------+-------------------------+-----------------------------------------------+
|   5 | event_scheduler | localhost | NULL   | Daemon  | 432662 | Waiting on empty queue  | NULL                                          |
| 132 | root            | localhost | sakila | Query   |    632 | User sleep              | select * from actor where actor_id=sleep(100) |
| 131 | root            | localhost | NULL   | Query   |    628 | Waiting for table flush | flush table with read lock                    |
+-----+-----------------+-----------+--------+---------+--------+-------------------------+-----------------------------------------------+
3 rows in set (0.00 sec)

mysql> 

mysql> kill 132;
Query OK, 0 rows affected (0.00 sec)

It can also be in sys.processlist、INFORMATION_SCHEMA.PROCESSLIST perhaps performance_schema.threads Found in it for a long time SQL, Just kill them .

modify xtrabackup Parameters of

For long execution time SQL, It can also be solved by modifying relevant parameters ：

• Use –ftwrl-wait-* Series parameters , It means ready for execution FLUSH TABLES WITH READ LOCK When encountering a long wait SQL Wait for the set time while executing the statement , Timeout has not been completed , Backup exit announced failure .
• Use –kill-long-query-* Series parameters , It means to encounter a long wait SQL Statement execution time , Overtime is kill Off execution SQL sentence , And you can set SQL Type defaults to all .
• Use –safe-slave-backup-* Series parameters , It means backing up from the Library , Stop the backup first slave sql Threads , Prevent excessive length SQL Statement execution , Start after backup slave sql Threads .

At the bottom of the article is my wechat , Welcome to join me ( toeic 890 Points of Oracle ACE).