Bug #74542 FLUSH TABLES WITH READ LOCK blocks EXPLAIN UPDATE/DELETE/INSERT

• Submitted: 24 Oct 2014 10:05
• Modified: 24 Oct 2014 10:50
• Reporter: Alexey Kopytov
• Status: Verified
• Category: MySQL Server: Optimizer
• Severity: S3 (Non-critical)
• Version: 5.6/5.7
• OS: Any
• CPU Architecture: Any

[24 Oct 2014 10:05] Alexey Kopytov

Description:
The global read lock acquired by FTWRL blocks EXPLAIN UPDATE/DELETE/INSERT even though technically there is no reason to do so.

mysql1> create table t (a int);
Query OK, 0 rows affected (0.02 sec)

mysql1> insert into t values(1);
Query OK, 1 row affected (0.01 sec)

mysql2> flush tables with read lock;
Query OK, 0 rows affected (0.00 sec)

mysql1> set lock_wait_timeout=1;
Query OK, 0 rows affected (0.00 sec)

mysql1> explain insert into t values(2);
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction
mysql1> explain delete from t;
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction
mysql1> explain update t set a=2;
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction

How to repeat:
See above.

[24 Oct 2014 10:50] MySQL Verification Team

Thank you for the bug report.