Today I encountered an interesting scenario. Upon failover of an AlwaysOn Availability Group, the secondary went from a resolving state and then abruptly the SQL service shut down. Obviously, the AG did not automatically come online.
It turns out that there is nothing wrong with SQL Server, but rather, the cause was a known bug in the windows server failover clustering of Windows 2008 R2. A patch is available. The KB is shown in the link below:
http://support.microsoft.com/kb/2777201
This was fixed with Windows 2012 and newer operating systems.
-Ken