[SQLObject] Antwort:Re: Creating more sqlobjects in _create()

[<ber...@zk...>]

Hi all

I think I found the cause of my deadlock, which is (if I'm right that i=
s),
located in main.py:912:

        val =3D cache.get(id, cls)
        if val is None:
            try:
                val =3D cls(_SO_fetch_no_create=3D1)

We reach this place e.g. by resolving a foreign key through
_SO_foreignKey() and get().

If I read the code correctly, the cache's lock is acquired in cache.get=
()
(but not released, which would be done in cache.finishPut(cls)), if the=

object doesn't already exist. cls(_SO_fetch_no_create=3D1) then acquire=
s the
class' lock through the threadSafeMethod() wrapper in declarative.py.

On the other hand, if I'm creating an object of the same type, at exact=
ly
the time between the cache.get() and cls() calls, I'll be able to acqui=
re
the class' lock and then pend on the cache's lock, through threadSafe
__init__(), then _create(), then _SO_finishCreate(), cacheSet.created()=
,
CacheFactory.created() and finally CashFactory.cull().

This leaves us with the first thread having acquired the cache's lock w=
hile
pending for the class' lock, and the second thread having acquired the
class' lock and pending for the cache's lock -> deadlock. The situation=
 can
only occur, if the creation of the object triggers a cache.cull(). In a=
ny
other case, the cache isn't locked and thus there are no problems.

Does this all make sense?

Bernhard

sql...@li... schrieb am 05.07.2007
17:11:20:

> Hi Oleg
>
> sql...@li... schrieb am 05.07.
> 2007 11:31:56:
>
> > On Thu, Jul 05, 2007 at 11:22:36AM +0200, ber...@zk... wr=
ote:
> > > The accesses are made in different threads, to objects of the sam=
e
class
> > > (MyClass in my example). One locks while resolving a ForeignKey i=
n
> > > _SO_foreignKey(), one locks in _SO_finishCreate() and the third o=
ne
in the
> > > selectresult's __iter__().
> >
> >    Can you do an experiment - create objects of OtherClass outside =
of
> > _create() and assign them to foreign keys after the object has been=

fully
> > created?
> >
>
> I've tried to produce an isolated example of my problem. I'm not yet
> sure if everything's needed in here, I'm still trying to buil it furt=
her
down.
>
> Things I noticed so far:
>
>  - The _create seems not to have anything to do with it, as it locks
> even without one.
>  - A higher number of threads increase the possibility for deadlocks.=

>  - I need to have two classes, MyClass1 and MyClass2, in order for
> the deadlock to occur.
>
> To the code:
>  - On my machine, the code deadlocks after some 10s of objects create=
d.
>  - The threadframe module may be downloaded here: http://www.majid.
> info/mylos/stories/2004/06/10/threadframe.html. Maybe you have some
> clever debugger that can do that for you... :-)
>

___________________________________________________________________

Disclaimer:


Diese Mitteilung ist nur fuer die Empfaengerin / den Empfaenger bestimm=
t.

Fuer den Fall, dass sie von nichtberechtigten Personen empfangen wird,
bitten wir diese hoeflich, die Mitteilung an die ZKB zurueckzusenden un=
d
anschliessend die Mitteilung mit allen Anhaengen sowie allfaellige Kopi=
en
zu vernichten bzw. zu loeschen. Der Gebrauch der Information ist verbot=
en.


This message is intended only for the named recipient and may contain
confidential or privileged information.

If you have received it in error, please advise the sender by return e-=
mail
and delete this message and any attachments. Any unauthorised use or
dissemination of this information is strictly prohibited.=