Hi Martin,
Does it mean that the returned session cookie should be propagated by the
client into all subsequent requests? Is it something that can be safely
assumed about AJAX-clients? Do they usually do it in typical cases? I'm
asking because I'm not a big expert in such client-side things.
In principle, as described in the initial post, there is a process id, which
is a unique id that could be used for correlation. The problem is that
currently it is just a part of initial reponse and later used by the client
in request URIs and not as an HTTP header or a session cookie.
-Leo
Martin Matula-3 wrote:
>
> Hi Leo,
>
> On 25.2.2011 10:16, Leo Romanoff wrote:
>> Now, I'd like to make the Jersey-based front-end for this service also
>> run
>> in the cluster mode. To achieve that I'd like to use load-balancing of
>> incoming startSession requests between nodes to produce an even load on
>> cluster members. And I need to achieve that all subsequent
>> readSessionState
>> and stopSession requests land at the right cluster node, that was picked
>> for
>> the initial startSession request.
> Depending on your load-balancer, it should be capable of providing the
> stickyness based on the session cookie.
> Martin
>
>
>
--
View this message in context: http://jersey.576304.n2.nabble.com/Providing-a-stateful-JAX-RS-service-in-a-cluster-tp6063881p6064086.html
Sent from the Jersey mailing list archive at Nabble.com.