[MDEV-6226] tokudb_lock_timeout 0 is not working as expected in MariaDB compared to Tokutek's version Created: 2014-05-10  Updated: 2014-05-12  Due: 2014-05-26  Resolved: 2014-05-12

Status: Closed
Project: MariaDB Server
Component/s: None
Affects Version/s: 10.0.10
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Alexander Loginov Assignee: Elena Stepanova
Resolution: Not a Bug Votes: 0
Labels: tokudb, tokudb_lock_timeout
Environment:

Debian Wheezy x64


 Description   

Hi.

From Tokutek documentation http://www.tokutek.com/wp-content/uploads/2013/10/mariadb-5.5.30-tokudb-7.1.0-users-guide.pdf:

4.2.5 tokudb_lock_timeout
....
A value of 0 disables lock waits
....

If tokudb_lock_timeout = 0, then lock waits should not happen. However with MariaDB 10.0.10 i've see, that if i request a table lock, the request refused with "#1205 - Lock wait timeout exceeded; try restarting transaction", so MariaDB now actually set lock timeout to zero milliseconds.

I think it will be better to match the behaviour with Tokutek version and disable lock waits, if tokudb_lock_timeout = 0.

Sincerely,

Alex.

 Comments   
Comment by Elena Stepanova [ 2014-05-12 ]

Hi Alexander,

What do you expect to happen when lock waits are disabled?
From what I see, the Tokutek version behaves exactly the same way: 

MySQL [test]> set tokudb_lock_timeout=0;
 
Query OK, 0 rows affected (0.00 sec)
 
 
 
MySQL [test]> update t1 set i = 2;
 
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction
 
MySQL [test]> select @@version;
 
+---------------------+
 
| @@version           |
 
+---------------------+
 
| 5.5.36-tokudb-7.1.5 |
 
+---------------------+
 
1 row in set (0.00 sec)
 
 
 
MySQL [test]> select @@version_comment;
 
+------------------------------+
 
| @@version_comment            |
 
+------------------------------+
 
| MySQL Community Server (GPL) |
 
+------------------------------+
 
1 row in set (0.00 sec)

and it's the only reasonable way I can think of – 0 means that waiting is disabled, not the timeout; it doesn't mean that the statement will actually break the lock of another transaction. If waiting is disabled, the timeout happens right away.

Comment by Alexander Loginov [ 2014-05-12 ]

Oh, it seems like i misunderstand the word.

Issue can be closed.

Generated at Thu Feb 08 07:10:18 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.