There should be no tweaking to get it working correctly (ruleflows don't
even use separate threads during execution currently). Based on your
explanation however, it is not obvious why things might be going wrong. I
suggest trying to use the audit view or a custom logger to figure out what
is going on in both cases. If you are losing records, there must be some
reason why the rules are not being executed for those records (e.g. the
record does not match the conditions of the rule, the ruleflow-group is
deactivated before all records have been processed, etc.). Try making a
small use case where you enter a number of records and not all records seem
to be processed. Then try looking at the audit log to see which rules are
activated and deactivated, which ruleflow-groups are activated and
deactivated, etc. If you do find a problem where you thing the ruleflow is
not behaving as expected, it would be nice if you could send me a
self-countained example that shows the problem and I'll try to look into it
asap.
One additional piece of advice: it does seem a little strange however that
you are using 50 ruleflow-groups for 116 rules. Ruleflow-groups are used to
specify the order in which large rule sets should be executed. If you want
to specify the order for single rules, other features like salience, agenda
groups, etc. might also be useful. But nevertheless, ruleflows should even
work in these circumstances.
Kris
----- Original Message -----
From: "Mazrim" <UE011434(a)guest.telecomitalia.it>
To: <rules-users(a)lists.jboss.org>
Subject: [rules-users] ruleflow trouble when process a large number of
records
Hello,
I'm using drools 4.0.4 on jboss 4.0.5.
I build a ruleflow and a set of rules following the documentation on
official drools site.
The ruleflow consist of about 50 ruleflow-group and 116 rules.
The ruleflow works correctly when processing a little number of records
(100-1000).
Instead if the number of records is around 20000 (they are processed in
sets
of 100 record) many rules don't fire; so many records are lost in output.
If I process in sets of 50 records I lose less records than in sets of 100
records. And If I process in sets of 20 records I lose less records than
in
sets of 50 records.
Must I tweak the JVM memory or the number of JVM threads? Or does this
issue
concern the building of ruleflow (maybe it's bad built)?
Thanks in advance
--
View this message in context:
http://www.nabble.com/ruleflow-trouble-when-process-a-large-number-of-rec...
Sent from the drools - user mailing list archive at
Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users