Core/DBLayer: Fix another possible crash on shutdown. This should be the last of them caused by this area of the core.

src/server/shared/Database/DatabaseWorker.cpp

@@ -47,6 +47,5 @@ int DatabaseWorker::svc()
         delete request;
     }
 
-    m_conn->Close();
     return 0;
 }

src/server/shared/Database/DatabaseWorkerPool.h

@@ -109,9 +109,9 @@ class DatabaseWorkerPool
                 /// Now we just wait until m_queue gives the signal to the worker threads to stop
                 T* t = m_connections[IDX_ASYNC][i];
                 DatabaseWorker* worker = t->m_worker;
-                worker->wait(); // t->Close(); is called from worker thread
+                worker->wait();
                 delete worker;
-                --m_connectionCount[IDX_ASYNC];
+                t->Close();
             }
 
             sLog->outSQLDriver("Asynchronous connections on databasepool '%s' terminated. Proceeding with synchronous connections.", m_connectionInfo.database.c_str());
@@ -123,7 +123,6 @@ class DatabaseWorkerPool
                 //while (1)
                 //    if (t->LockIfReady()) -- For some reason deadlocks us
                 t->Close();
-                --m_connectionCount[IDX_SYNCH];
             }
 
             sLog->outSQLDriver("All connections on databasepool %s closed.", m_connectionInfo.database.c_str());