[rules-users] Drools 5.1 Flow in a "Clustered" environment

[Kris Verlaenen]

Ken,

In Drools, timers etc. are linked to a session (where the session can 
contain process instances, timers, data, rule evaluations, etc.). So, 
even if you are using Drools Flow in a clustered environment, where you 
could have a number of independent sessions running simultaneously, the 
timer will always be associated to that session. There are two possible 
strategies then: either (1) you keep the session online for ever (your 
cluster has a number of running sessions and you're possibly using 
persistence to restore in case of system failure), and the timers will 
fire in that session only when the time is right or (2) you only keep 
your session alive when it is necessary so sessions can go offline, 
where you could then use a scheduler to make sure the session is brought 
back online (from persistent storage) whenever a timer should fire in an 
offline session. The latter is currently being implemented as a 
community contribution.

Kris

Ken Young wrote:
> We are looking at embedding Drools 5.1 (specifically Drools Flow) in our application, but I am coming up short on determining how to configure clustering, or the support.  
>
> Specifically, I want to ensure that background processes that are taking place (reminders, timers, etc.) only happen once on a cluster.  Does drools support this and how?  
>
> In other software that we have clustered (Quartz for example), where a shared database lock is used to ensure the the processing happens in one place.  I was wondering if Drools 5.1 worked in a similar fashion.
>
> Thanks
> Ken
>
>
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>