I think I have found the problem! I took a closer look at the Fatal error core:

#0  0xb76fe424 in __kernel_vsyscall ()
#1  0xb740ccb1 in *__GI_raise (sig=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
#2  0xb740e3f2 in *__GI_abort () at abort.c:92
#3  0xb766193f in Py_FatalError (msg=0xb76949de "GC object already tracked")
    at Python/pythonrun.c:1670
#4  0xb7595a3a in PyMethod_New (func=<function at remote 0x8d649cc>, self=
    <QtReactor(waker=<_UnixWaker(i=13, fileno=<function at remote 0x93f6994>, reactor=<...>, o=14) at remote 0x93f3dac>, threadCallQueue=[(<instancemethod at remote 0x9b5639c>, (0,), {})], _cancellations=0, _justStopped=False, _newTimedCalls=[], _pendingTimedCalls=[<DelayedCall(resetter=<instancemethod at remote 0x9b5d504>, seconds=<built-in function time>, args=(), canceller=<instancemethod at remote 0x9cb48ec>, delayed_time=0, kw={}, func=<LoopingCall(a=(...), interval=2, clock=<...>, _expectNextCallAt=<float at remote 0x9ab7b44>, f=<instancemethod at remote 0x96db34c>, deferred=<Deferred(_canceller=None, callbacks=[((<function at remote 0x8bc6f44>, None, None), (<instancemethod at remote 0x96db374>, (...), {}))]) at remote 0x96da0cc>, running=True, kw={}, starttime=<float at remote 0x8c3f0f4>, call=<...>, _runAtStart=True) at remote 0x96d776c>, time=<float at remote 0x9d4cb74>, cancelled=0, called=0) at remote 0x9d1184c>, <DelayedCall(resetter=<instancemethod at remote 0xa2c4edc>, seconds=<built-in function ti...(truncated), klass=<type at remote 0x8e5a3dc>) at Objects/classobject.c:2250
#5  0xb6c71225 in PySide::DynamicSlotDataV2::callback (this=0x96b6440)
    at /var/tmp/portage/dev-python/pyside-1.0.9/work/pyside-qt4.7+1.0.9/libpyside/globalreceiverv2.cpp:129
#6  0xb6c718fc in PySide::GlobalReceiverV2::qt_metacall (this=0x9519b88, 
    call=QMetaObject::InvokeMetaMethod, id=5, args=0xbffbeef4)
    at /var/tmp/portage/dev-python/pyside-1.0.9/work/pyside-qt4.7+1.0.9/libpyside/globalreceiverv2.cpp:289
#7  0xb6ab76a7 in QMetaObject::metacall (object=0x9519b88, cl=QMetaObject::InvokeMetaMethod, 
    idx=5, argv=0xbffbeef4) at kernel/qmetaobject.cpp:237
#8  0xb6ac9b8c in QMetaObject::activate (sender=0x9760e40, m=0xb6c5d644, 
    local_signal_index=<optimized out>, argv=<optimized out>) at kernel/qobject.cpp:3278
#9  0xb6b21b27 in QTimer::timeout (this=0x9760e40) at .moc/debug-shared/moc_qtimer.cpp:128
#10 0xb6acfb01 in QTimer::timerEvent (this=0x9760e40, e=0xbffbf41c) at kernel/qtimer.cpp:271
#11 0xb6e35efb in QTimerWrapper::timerEvent (this=0x9760e40, arg__1=0xbffbf41c)
    at /var/tmp/portage/dev-python/pyside-1.0.9/work/pyside-1.0.9_build/PySide/QtCore/PySide/QtCore/qtimer_wrapper.cpp:233
#12 0xb6ac2e67 in QObject::event (this=0x9760e40, e=0x6) at kernel/qobject.cpp:1181
#13 0xb6e37717 in QTimerWrapper::event (this=0x9760e40, arg__1=0xbffbf41c)
    at /var/tmp/portage/dev-python/pyside-1.0.9/work/pyside-1.0.9_build/PySide/QtCore/PySide/QtCore/qtimer_wrapper.cpp:164


If this would explain the update_refs lock up then I will close this bug... (i also opening on on pyside)