MySQL Bugs: #19998: SET GLOBAL wait_timeout is not propagated to new connections

Bug #19998	SET GLOBAL wait_timeout is not propagated to new connections
Submitted:	22 May 2006 11:56	Modified:	22 May 2006 16:26
Reporter:	Jan Kneschke	Email Updates:
Status:	Not a Bug	Impact on me:	None
Category:	MySQL Server	Severity:	S3 (Non-critical)
Version:	5.0.16, 5.1-bk	OS:	Linux (Linux/x86, MacOSX)
Assigned to:		CPU Architecture:	Any

Description:
the wait_timeout for new connections can't be changed at runtime.

SET GLOBAL wait_timeout = 12000 is setting the global wait_timeout as expected, but if you reconnect with the mysql commandline client it still has to old wait_timeout (28800) set.

How to repeat:
Using the MySQL client:

> show variables like 'wait_timeout';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| wait_timeout  | 28800 |
+---------------+-------+

> set global wait_timeout = 12000;
Query OK, 0 rows affected (0.00 sec)

> show global variables like 'wait_timeout';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| wait_timeout  | 12000 |
+---------------+-------+

> quit
... login again ...

> show global variables like 'wait_timeout';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| wait_timeout  | 12000 |
+---------------+-------+
1 row in set (0.00 sec)

> show variables like 'wait_timeout';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| wait_timeout  | 28800 |
+---------------+-------+
1 row in set (0.00 sec)

> select @@wait_timeout;
+----------------+
| @@wait_timeout |
+----------------+
|          28800 |
+----------------+
1 row in set (0.00 sec)

Suggested fix:
propagate wait_timeout from the global settings to the session settings for new connections.

Verified at fresh 5.1-bk build

Did you remember to change interactive_timeout? (That's where the setting for a session's wait_timeout comes from when it has been flagged as an interactive connection, which the mysql client does.)

See http://dev.mysql.com/doc/refman/4.1/en/server-system-variables.html

This bug should not have been marked 'Verified' without checking this, since it is not mentioned in the bug report.

Right, this "feature" is working as documented in the manual.

For a user this a non-obvious and from his perspective still a bug, even if it is documented.

There are many different things not obvious to users and even support engineers or developers. Not all of them are bugs, though. So, this one looks like intended and documented behavior - not a bug.