Seperate JVM running clustered Inifinispan on 4.2 EAP

[blieb73]

We are currently running EAP 4.2 on 4 non clustered instances.  We are using jboss cache independently right now, but we have a pressing need to have a shared cache for data consistency.

I have searched the forums and found the jgroups jar would cause some issues for me running infinispan in the same jvm as the application.  I also read about the "hotrod" server that allows Infinispan to run in its own space.

What I'm looking for is a way to cluster the hotrod servers (I would run one on each server corresponding to the application) and have jboss use that as its second level cache.

Is this too far fetched?  The alternative is to upgrade to EAP 5 and cluster our instances, but that is going to take a lot more work.

[sannegrinovero]

Hello Benjamin,

before we start discussing all Infinispan options, did you consider using the existing JBoss Cache for the same purpose?

As second level cache you mean your goal is to use it as cache for Hibernate entities?

Also, EAP5 is not using Infinispan either so that won't change your scenario.

Since you're using EAP I'd suggest you to open a support ticket; don't expect support to provide Infinispan for 4.2 as that's not possible, but at least the business area will track this request, and push and or speedup for some solution.

Hotrod might be a good option, but it depends from your requirements; are you running Java6 at least? Infinispan requires Java6, I'm not sure if hotrod would work on Java5 as we build that with 6 too, but we could investigate that.

[sannegrinovero]

almost forgot: the next EAP including Infinispan is EAP6, based on the community edition AS7. They are both going to be released pretty soon, nad we're very excited about how good they are so there might be more arguments - on top of Infinispan - to upgrade to it. I'd suggest to download a beta of AS7 to get an idea, you won't be disappointed!

[blieb73]

Sanne, thanks for the response.  I am using java6 with EAP 4.2.0.  I would like to use either one hotrod server or multiple clustered hotrod servers to act as the hibernate second level cache for four non clustered machines.  We currently use jboss cache, but the caches are separate. 

I say hotrod because it may be the least invasive (and maybe the only option for 4.2).  Pretty sure I have to get hibernate up to 3.5 and include the infinispan jar requirements for this to work.  I would then use the hotrod server via jndi in my xml file.  Am I on the right track here?