that does seem odd. Are the system clocks in sync on both machines? It would be interesting to get db logs to see the exact sequence of statements with values. The idea is that a lease update would fail if the slave takes over, but that does not seem to be the case on your master. It appears that the slave is checking for expiry about the same time that the master should be updating, so it becomes a master, but in this case the slave should give up.
Consider using the same timeout values on both the master and slave but it does look like there is a bug here. On 13 December 2012 07:10, mandar.wanpal <mandar.wan...@gmail.com> wrote: > lockAcquireSleepInterval -- http://redhat.com http://blog.garytully.com
