MySQL Bugs: #57719: Database clock stops

Bug #57719	Database clock stops
Submitted:	25 Oct 2010 19:36	Modified:	10 Nov 2010 23:21
Reporter:	chris valaas	Email Updates:
Status:	Duplicate	Impact on me:	None
Category:	MySQL Server: General	Severity:	S2 (Serious)
Version:	5.4.* 5.5.6	OS:	Solaris
Assigned to:		CPU Architecture:	Any

Description:
Periodically, mysqld's clock stops. The system's clock is fine. InnoDB's clock actually seems fine, too. 
Here's a log of the shutdown after we noticed the problem:

101025 11:01:27 [Note] /usr/local/mysql/bin/mysqld: Normal shutdown
101025 11:01:27 [Note] Event Scheduler: Purging the queue. 0 events
101025 11:38:28  InnoDB: Starting shutdown...
101025 11:48:55  InnoDB: Shutdown completed; log sequence number 16128739147474
101025 11:01:27 [Note] /usr/local/mysql/bin/mysqld: Shutdown complete
101025 11:48:59 mysqld_safe mysqld from pid file /u0/mysql_data/mike.pid ended

The InnoDB timestamps match the system's clock, while the main mysqld timestamp is frozen at 11:01:27

We had this problem on version 5.4 and now with 5.5.6, both 64-bit, both precompiled binaries.

This seems to only happen on our master server, a Sun (Oracle?) x4600. Several of our slaves are also x4600s, but have not experienced this issue.

How to repeat:
Run mysqld for awhile and wait for it to happen. Seems to be about once per month for us.

Here are the last few times it's happened:

10/25/10 - 11:02a
09/29/10 - 01:25a
07/21/10 - 07:18a
06/06/10 - 03:43a

have been a few reports of this, especially in virtualized environments like vmware.  does SELECT NOW() also return the same time ?

SELECT NOW() also returns the "frozen" time. And we are not using any virtualization.

Please do not submit the same bug more than once. An existing bug report already describes this very problem. Even if you feel that your issue is somewhat different, the resolution is likely
to be the same. Because of this, we hope you add your comments to the original bug instead.

Thank you for your interest in MySQL.

Duplicate of bug #42054