What could a Database reboot be doing to fix a hanging query

If you have a query that is stuck in processing, what would a database reboot be doing that would fix the query after restarting Statistics don't appear to be involved here as they are updated and are not changed before, and after, the reboot. This has happned 3 times where a reboot of the database fixed the problem.

We are using MSSQL 2008 ( only version supported by the application) on WIndows server 2008

I have query that that can hang will , only at times, but not all the time  ( This is inside a cobol program)  When the database is rebooted, the query completes in 10 minutes. What would the reboot be doing to fix the problem with the query.  Here are the symptoms of  of the query duing the hanging process and what was done up to this point.

Symtoms: Numbered 1-6
When running this statement
1. SELECT session_id,status,row_count,reads,writes,logical_reads,cpu_time FROM sys.dm_exec_sessions for that SPID, (all numbers run to a certain point and then freeze.

2. When the TEMPDB is examined , the numbers freeze and CPU time is stopped
FROM sys.dm_db_file_space_usage;
SELECT
                    sys.dm_exec_sessions.session_id AS [SESSION ID],
                    DB_NAME(database_id) AS [DATABASE Name],
                    HOST_NAME AS [System Name],
                    program_name AS [Program Name],
                    login_name AS [USER Name],
                    status,
                    cpu_time AS [CPU TIME (in milisec)],
                    total_scheduled_time AS [Total Scheduled TIME (in milisec)],
                    total_elapsed_time AS    [Elapsed TIME (in milisec)],
                    (memory_usage * 8)      AS [Memory USAGE (in KB)],
                    (user_objects_alloc_page_count * 8) AS [SPACE Allocated FOR USER Objects (in KB)],
                    (user_objects_dealloc_page_count * 8) AS [SPACE Deallocated FOR USER Objects (in KB)],
                    (internal_objects_alloc_page_count * 8) AS [SPACE Allocated FOR Internal Objects (in KB)],
                    (internal_objects_dealloc_page_count * 8) AS [SPACE Deallocated FOR Internal Objects (in KB)],
                    CASE is_user_process
                                         WHEN 1      THEN 'user session'
                                         WHEN 0      THEN 'system session'
                    END         AS [SESSION Type], row_count AS [ROW COUNT]
FROM sys.dm_db_session_space_usage INNER join sys.dm_exec_sessions
     ON sys.dm_db_session_space_usage.session_id = sys.dm_exec_sessions.session_id
3. dbcc opentran shows the LSN has stopped moving for the SPID and but continues to show the transaction running
4. The query does show locks on the tables (sp_lock2) but the locks are exlusive to that SPID
5. When the database is refreshed to another instance, the process runs fine.
6. Several times, when this happened, the query was trapped in the DBCC inputbuffer( spid ) and analyzed.  statistics were updated after the hang, but it was only the  reboot still fixed the problem.

Other things tried at this point.
Statistics were updated on all tables that are involved in the query trapped by the DBCC Input buffer.
Before the reboot, the DBCC FREEPROCCACHE was executed and the attempt was made again, still the query hangs in a "running" state.