Hi again,

Guillaume Nodet wrote:
Ok, sorry !
I've rewritten the repository.xml beacause i use my
own generator to build it and the object-cache tag had
vanished...

ok, user failure ;-)

However, i've got another question about object deletion.
I'm currently using the deleteByQuery method got two problems with it:
first, objects are not removed from the cache, and second it does
not seam to perform sub-objets deletion (aka collection descriptors).
I think if i call getCollectionByQuery and iterate through
the collection to call delete on each object, these two problems will
disappear, but is this the intended behavior of the deleteByQuery method ?


you are right, this is an open issue (or not adequate documented). I discuss this many weeks ago with Jakob. The problem is object materialization. If we materialize all objects before delete, we could run into memory problems for huge delete operations. If we don't do this, we never can remove auto-delete references and cache entries.


As far as I know Jakob thought about to materialize the Identity objects only, but I'm not sure. Recommend to discuss this issue on the dev-list.

regards,
Armin

-----Message d'origine-----
De : Armin Waibel [mailto:[EMAIL PROTECTED]
Envoye : vendredi 30 janvier 2004 10:56
A : OJB Users List
Objet : Re: Optimistic locking error with SequenceManagerHighLowImpl


Hi Guillaume,


Guillaume Nodet wrote:


When using two instances of ojb on different data sets, but using the same
table for SequenceManagerHighLowImpl, i ran into an optimistic lock error.
I took a look at the code, and i see there were 5 retries to test for
achieving
the select / store process.
My question is: as the HighLowSequence is cached by the cache
implementation, shouldn't it be necessary to remove it from the cache

before


retrying the process ?
I suppose that the cache returns the same object every times it is queried
so the
process will fail at each new attempt.



Which version of OJB do you use? Normally in repository_internal.xml of
HighLowSequence class-descriptor the empty cache implementation was
declared, thus objects should never be cached.

...
<class-descriptor
          class="org.apache.ojb.broker.util.sequence.HighLowSequence"
          table="OJB_HL_SEQ"
    >

        <object-cache
class="org.apache.ojb.broker.cache.ObjectCacheEmptyImpl">
        </object-cache>
...

regards,
Armin


Regards,

Guillaume Nodet



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]







---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]





---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]






--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]



Reply via email to