A serious deadlock with 7u45, it is all a timing thing between different threads.  It seems there is some lock contention in the sun.security.ssl.SSLSocketImpl. Because it is a threading problem, it does not occur very predictably and have not been able to reproduce it yet on the test computer.

However, attached thread dump got from the VM after the deadlock shows the locks that each thread has and where the deadlock occurred.  These are generally caused by different execution paths obtaining the same locks in different orders.   Since the two execution paths are represented in the attachment, this will be sufficient to analyze the situation.