Beginners Clustering Qu

[hay7777]

Hi,

I asked this on the beginners list, but then realised there was a seperate Clustering forum!

I am a beginner with JBoss, J2EE and with clustering, but proficient in Java, servlets and JSP.

We are currently using Tomcat as our webserver. On the back end we have our own server (not a webserver) that we submit jobs to for it to process. So we have our clients talking to the webserver, and the webserver submitting jobs to our custom server behind it.

We need to expand things to an enterprise level, and have multiple back-end servers which should be load balanced/fail over etc..

Can JBoss's clustering help in this situation ie specifically, helping with the load balancing for the back-end servers, security for the whole system, transactions etc.?

many thanks,

David

[anil.saldhana]

Yes. It can do all the things you asked for. Look on the wiki for the relevant pages to get more inisight into the Jboss power.

[hay7777]

Hi. Okay...looked at the Wiki, but can't get an overall picture of how this will work.

We have the following setup:

clients -> web server -> job handling server

We want to cluster both the web server, and the server that we submit jobs to.

Can anyone give me a high-level glimpse of how JBoss would help in this situation?

Many thanks,


David

[lretief]

Does the Job Handling Server run on JBoss? From your initial mail I am guessing no? Is it third-party, custom, Java whatever?

[hay7777]

We are trying to figure out how to set things up, so, if it's feasible, we could run it within JBoss if it would help.

As a beginner to all this - what determines if it runs "within" JBoss - on the same server? It's written in C++, so we connect to it at the moment from our jave code running as a webapp on Tomcat by opening a socket to it.

You can tell I am very new to all this!

A related question which I posted - if this Job Handling Server is not running within JBoss, but on the same machine, and the Job Handling Server stops responding, is there any way to mark that machine as not in the cluster any more?

Many thanks for any help!!

cheers,

David

[starksm64]

You would not want to mark the node as down, just that this node no longer has the service replicated to it. This would be handled creating an HA service facade for the job handler, and deploying/undeploying the service based on the availability of the job handler.

[hay7777]

Hi Scott,

Thanks for the reply!

I'm sure your busy, but would you mind explaining a little more - I'm very new to JBoss, and not sure how this would work?

thanks!

David