[rules-users] java.lang.ClassCastException: org.drools.reteoo.BetaMemory cannot be cast to org.drools.reteoo.AlphaNode$AlphaMemory
dcrissman
dcrissman at redhat.com
Mon May 6 15:50:26 EDT 2013
It wasn't a rule issue. I was effectively re-adding a change set to the
knowledge base multiple times. So *IF* there is an issue here, it would only
maybe be some kind of error/warning stating that a rule has already been
added to the kb.
Davide Sottara wrote
> The Stateless Session wraps a Stateful one.. it would be extremely
> useful if you could narrow down your initial KB and your changeset
> to the minimal set of rules that causes this corruption. There could
> be a bug in the incremental construction process.
> Thanks
> Davide
>
> On 05/06/2013 11:33 AM, Wolfgang Laun wrote:
>> Sorry, but I doubt that anyone can benefit from this somewhat
>> obscure description of a use case - but thanks for trying.
>> -W
>>
>> On 06/05/2013, dcrissman <
> dcrissman@
> > wrote:
>>> While working on a sanitized version of the code, I discovered my issue.
>>>
>>> I am using a StatefulKnowledgeSession essentially statelessly. I am
>>> doing
>>> this because in the past I have found StatefulKnowledgeSession to be
>>> more
>>> feature rich. Regardless, each execution created a new
>>> StatelessKnowledgeSession instance which is built from a single instance
>>> of
>>> a KnowledgeBase. During the setup for each run, I was adding my
>>> change_set
>>> to that KnowledgeBase, causing the error I was getting. I'd like to
>>> think
>>> that my mistake could be recycled and benefit someone else in the
>>> future.
>>>
>>> Davide - Thank you for your time. Sorry that this turned out to be user
>>> error.
>>>
>>> -Dennis
>>>
>>>
>>> Davide Sottara wrote
>>>> Could you post **at least** the unit test throwing the exception,
>>>> the one that "enables" it and the code you are using to run them?
>>>> This may be yet another blocker for 5.6, so I'd like to investigate it
>>>> as soon as possible
>>>> Thanks!
>>>> Davide
>>>>
>>>> On 05/02/2013 07:20 AM, dcrissman wrote:
>>>>> Yes, I am getting this exception (sadly, not something that I am just
>>>>> worried
>>>>> about).
>>>>>
>>>>> I am checking for rule complication errors before running the rules,
>>>>> and
>>>>> no
>>>>> errors are being given.
>>>>>
>>>>> I am not dynamically adding or removing rules once I have my knowledge
>>>>> base
>>>>> setup, it does not change though the duration of the run.
>>>>>
>>>>> I am using a StatefulKnowledgeSession, and I make sure that after each
>>>>> run
>>>>> that I am calling dispose in a finally block. So shouldn't be any
>>>>> strange
>>>>> memory issues.
>>>>>
>>>>> To add to the original description, I am getting these exceptions in
>>>>> my
>>>>> unit
>>>>> tests. There is one unit test suite that when disabled the exception
>>>>> goes
>>>>> away, but when enabled the exceptions happen in other suites. Which of
>>>>> course is odd and led me to believe that there is a problem with the
>>>>> unit
>>>>> tests (which is still not ruled out). But the suites are using
>>>>> different
>>>>> sessions, same rules, but totally different instances of everything.
>>>>>
>>>>> The only thing I am doing that may be a little funny is consolidating
>>>>> a
>>>>> collection of objects into one then retracting the collection, but
>>>>> when
>>>>> I
>>>>> comment out this rule I still am getting this exception. Aside from
>>>>> this
>>>>> one
>>>>> case, I am not doing anything crazy.
>>>>>
>>>>> [code]
>>>>> //ArrayList provided by drools seems to dynamically adjust with the
>>>>> retractions, this safeguards us against that.
>>>>> ArrayList safetyArr = new ArrayList($attrs);
>>>>>
>>>>> ArrayList values = new ArrayList();
>>>>> for (int i=0; i < safetyArr.size(); i++){
>>>>> Attribute attr = ((Attribute) safetyArr.get(i));
>>>>> values.add(attr.getValue());
>>>>> retract(attr);
>>>>> }
>>>>>
>>>>> insert(newAttribute($redHatName, StringUtils.join(values, ",")));
>>>>> [/code]
>>>>>
>>>>> Any other thoughts you may have would be appreciated.
>>>>>
>>>>> Thanks!
>>>>>
>>>>>
>>>>> Davide Sottara wrote
>>>>>> The original thread was for a bug in 5.0.x, fixed in 5.1, but it
>>>>>> would
>>>>>> be
>>>>>> a
>>>>>> very serious issue in 5.5. It basically means that the local memory
>>>>>> of
>>>>>> a
>>>>>> beta node has been assigned to an alpha node, or that the wrong
>>>>>> memory
>>>>>> had
>>>>>> been retrieved. In any case, the rete is broken: wrong facts or types
>>>>>> would
>>>>>> be joined in a different order or way than stated by the rules.
>>>>>>
>>>>>> Is this something you're experiencing , or just worried that it may
>>>>>> happen?
>>>>>>
>>>>>> Causes could be an invalid rule in the first place (running the
>>>>>> engine
>>>>>> without checking for compilation errors), a bug in the construction
>>>>>> process
>>>>>> (unlikely, unless the rule is a very rare edge case), or a complex
>>>>>> scenario
>>>>>> where you add and remove rules dynamically from a knowledge base.
>>>>>> On May 1, 2013 2:25 PM, "dcrissman" <
>>>>>> dcrissman@
>>>>>> > wrote:
>>>>>>
>>>>>>> Could you explain what you mean by rete corruption? I am familiar
>>>>>>> with
>>>>>>> rete,
>>>>>>> but hardly an expert. I am not really doing anything too
>>>>>>> complicated,
>>>>>>> I
>>>>>>> guess some runtime inserts and retracts.
>>>>>>>
>>>>>>> I tried 6.0.0.Beta2, but that has more api changes then I am
>>>>>>> prepared
>>>>>>> to
>>>>>>> handle, I'll wait until the Final version.
>>>>>>>
>>>>>>> As for 5.6.0-SNAPSHOT, I am having trouble pulling it down. I will
>>>>>>> pick
>>>>>>> that
>>>>>>> up tomorrow.
>>>>>>>
>>>>>>> In the meantime, any clarification to the rete question would be
>>>>>>> appreciated.
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Dennis
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> View this message in context:
>>>>>>> http://drools.46999.n3.nabble.com/java-lang-ClassCastException-org-drools-reteoo-BetaMemory-cannot-be-cast-to-org-drools-reteoo-AlphaNy-tp4023610p4023612.html
>>>>>>> Sent from the Drools: User forum mailing list archive at Nabble.com.
>>>>>>> _______________________________________________
>>>>>>> rules-users mailing list
>>>>>>>
>>>>>> rules-users at .jboss
>>>>>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>>>>>
>>>>>> _______________________________________________
>>>>>> rules-users mailing list
>>>>>> rules-users at .jboss
>>>>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> View this message in context:
>>>>> http://drools.46999.n3.nabble.com/java-lang-ClassCastException-org-drools-reteoo-BetaMemory-cannot-be-cast-to-org-drools-reteoo-AlphaNy-tp4023610p4023630.html
>>>>> Sent from the Drools: User forum mailing list archive at Nabble.com.
>>>>> _______________________________________________
>>>>> rules-users mailing list
>>>>>
>>>> rules-users at .jboss
>>>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>>>
>>>> _______________________________________________
>>>> rules-users mailing list
>>>> rules-users at .jboss
>>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>
>>>
>>>
>>>
>>> --
>>> View this message in context:
>>> http://drools.46999.n3.nabble.com/java-lang-ClassCastException-org-drools-reteoo-BetaMemory-cannot-be-cast-to-org-drools-reteoo-AlphaNy-tp4023610p4023693.html
>>> Sent from the Drools: User forum mailing list archive at Nabble.com.
>>> _______________________________________________
>>> rules-users mailing list
>>>
> rules-users at .jboss
>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>>
>> _______________________________________________
>> rules-users mailing list
>>
> rules-users at .jboss
>> https://lists.jboss.org/mailman/listinfo/rules-users
>>
>
> _______________________________________________
> rules-users mailing list
> rules-users at .jboss
> https://lists.jboss.org/mailman/listinfo/rules-users
--
View this message in context: http://drools.46999.n3.nabble.com/java-lang-ClassCastException-org-drools-reteoo-BetaMemory-cannot-be-cast-to-org-drools-reteoo-AlphaNy-tp4023610p4023697.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
More information about the rules-users
mailing list