[rules-users] Drools Fusion inconsistencies at increasing event throughputs

Mark Proctor mproctor at codehaus.org
Tue Apr 1 10:17:58 EDT 2014


We have this fixed now. Hopefully this will be in the nightly build, and available for testing tomorrow. I’ll ping with a link to the nightly builds once it’s available.

Mark

On 31 Mar 2014, at 14:34, Mark Proctor <mproctor at codehaus.org> wrote:

> We are still working on this, we’ve found a number of related points, as we are doing a full audit of the code. We’ll post as soon as we have the fixes available in a nightly build, so that people can test.
> 
> Mark
> On 29 Mar 2014, at 12:50, Mark Proctor <mproctor at codehaus.org> wrote:
> 
>> Mario has found an issue at a sync point, which can happen if it’s trying to schedule a timer at the same time that it’s firing that same timer from an update. We are trying to resolve that now. See lines 121 and 394, which later impacts lines 289 and 330.
>> https://github.com/droolsjbpm/drools/blob/master/drools-core/src/main/java/org/drools/core/phreak/PhreakTimerNode.java
>> 
>> Once we have this solved, we should have expected behaviour. Hopefully you can try this in the next 6.1 beta, in 2 weeks time.
>> 
>> Mark
>> On 29 Mar 2014, at 09:00, Vieri <vieri.emiliani at gmail.com> wrote:
>> 
>>> Mark,
>>> thanks for your update. 
>>> I don't know if this can help, but I added a simple AgendaEventListener to the test case and it seems that at some point drools stops matching the "Create event" rule, while facts (well, events) are still inserted and other rules are triggered properly.
>>> Vieri
>>> 
>>> 
>>> 
>>> On 28 March 2014 16:57, Mark Proctor <mproctor at codehaus.org> wrote:
>>> we’ve being auditing the code here, and we’ve found an issue on what happens when a rule is re-matched and the timer updated. Mario is addressing this now, and also re-viewing dropping of output. He’ll post with an update soon.
>>> 
>>> Mark
>>> On 25 Mar 2014, at 16:09, Vieri <vieri.emiliani at gmail.com> wrote:
>>> 
>>>> First of all, thanks for the support.
>>>> Mario, I confirm that the case you provided works fine also at my side. Still, have you tried my first example (using the cron directly to the counting rule)? Can you confirm it is not working, since this was my first concern?
>>>> As I said in the previous post, using the CronTrigger pattern greatly improves stability (I managed to run it at 500 eps), but it is not resolutive.
>>>> Adding a few rules, it's enough to go back to instability. 
>>>> 
>>>> 
>>> 
>>> _______________________________________________
>>> rules-users mailing list
>>> rules-users at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/rules-users
>>> 
>>> 
>>> 
>>> -- 
>>> Vieri Emiliani
>>> _______________________________________________
>>> rules-users mailing list
>>> rules-users at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/rules-users
>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/rules-users/attachments/20140401/f96e007f/attachment-0001.html 


More information about the rules-users mailing list