Details
-
Type:
Bug
-
Status: Closed
-
Priority:
Minor
-
Resolution: Fixed
-
Affects Version/s: 10.0.14
-
Fix Version/s: 10.0.15
-
Component/s: Storage Engine - InnoDB, Storage Engine - XtraDB
Description
http://bugs.mysql.com/bug.php?id=72123
lock_timeout thread works in a tight loop waking up every second
and checking for lock_wait_timeout. In addition, when a mysql
thread is forced to wait on a lock, it signals the lock_timeout thread
as well. This call is not required. In a heavily contended workload
each thread going to wait will signal the lock_timeout thread making
it work all the time. As lock_timeout thread scans the array of
waiting threads under lock_sys::wait_mutex which is already very
hot in contneded loads, these extra scans can cause significanct
performance regression.
Also, in various codepaths lock_timeout thread is signalled where
actual intention was to signal the innodb monitor thread.
Gliffy Diagrams
Attachments
Issue Links
- is part of
-
MDEV-6039 WebScaleSQL patches
-
- Closed
-
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
revno: 4455
committer: Jan Lindström <jplindst@mariadb.org>
branch nick: 10.0-innodb
timestamp: Fri 2014-10-24 17:56:04 +0300
message:
MDEV-6933: Spurious lock_wait_timeout_thread wakeup in lock_wait_suspend_thread()Merged Facebooks commit 6e06bbfa315ffb97d713dd6e672d6054036ddc21
authored by Inaam Rana from https://github.com/facebook/mysql-5.6.
Fixes MySQL bug http://bugs.mysql.com/bug.php?id=72123
lock_timeout thread works in a tight loop waking up every second
and checking for lock_wait_timeout. In addition, when a mysql
thread is forced to wait on a lock, it signals the lock_timeout thread
as well. This call is not required. In a heavily contended workload
each thread going to wait will signal the lock_timeout thread making
it work all the time. As lock_timeout thread scans the array of
waiting threads under lock_sys::wait_mutex which is already very
hot in contneded loads, these extra scans can cause significanct
performance regression.
Also, in various codepaths lock_timeout thread is signalled where
actual intention was to signal the innodb monitor thread.