Thanks Vladimir for the detailed article.
I have some doubts wrt failover: Let's say node N owned keys=a,b,c and a task T was
sent to N to work on a,b. Assuming N' was next in hash wheel to N, if N fails, task T
would be sent N' since this would be the new owner for a,b, wouldn't it? Note that
Infinispan uses fixed positions in hash wheel. I don't think there's a need here
to spawn T to all a,b backup nodes. Interaction between Infinispan F/J and rehashing will
be important here.
Another thing to consider is what to do if the T fails but not because node N failed, but
cos the task itself failed. You can either send the task to a,b backup owners which would
N' or consider T as failed directly.
Infinispan F/J will have to differentiate between these two scenarios.
On Jul 21, 2010, at 7:42 PM, Vladimir Blagojevic wrote:
Hi,
I wanted to gather feedback from community regarding proposed scope and objectives of
Distributed execution framework targeted for Infinispan 5.0. The document is a gentle
introduction to the subject of distributed execution and as such is accessible to wider
audience. Eager to hear your thoughts.
Regards,
Vladimir
https://community.jboss.org/wiki/InfinispanDistributedExecutionFramework
--
Vladimir Blagojevic
JBoss Clustering Team
JBoss, by Red Hat
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache