I've got a problem with an InnoDB database and I'm not really sure what setting to adjust to fix it. The problem I'm having is connections come in and start a transaction. Before the commit is issued the remote end then forcefully kills the connection (usually with a kill -9) and the rows just stay locked for hours and hours. I've tried setting the global wait_timeout to 3600 seconds but the rows still stay locked for 20,000+ seconds. Can someone point me in a direction of what I need to change? I don't have control over the remote server and I've already smacked the programmers for doing this. However, I need a DB side solution as they say it will take months to redesign that part while they are blaming the database for their issues.
↧