Entity Beans within different clusters. Problem ??

Discussions

EJB programming & troubleshooting: Entity Beans within different clusters. Problem ??

  1. Hello to all,

    Here is our problem. We are debating whether to use Entity Beans for persistence. We are using Websphere 4.0 and EJB 1.1.

    Now the problem is that we have 2 completly different clusters of App. Servers with the same application running on them. The clusters don't have any knowledge of each other. Now if we were to use CMP would that cause any data incosistencies ? or is there any other problems that could arise due to the use of Entity Beans with our current deployment structure ? Thank you.
  2. Do you mean two different application cluster that both talk to the same database? Entity beans won't cause you any particular difficulties, here. The EJB server is responsible for synchronizing entity data with the database, and therefore the data will be the same in both applications.

    Of course, you will have some limitations on your data caching options, but those will be a problem no matter what persistence technology you use.
  3. Do you mean two different application cluster that both talk to the same database? Entity beans won't cause you any particular difficulties, here. The EJB server is responsible for synchronizing entity data with the database, and therefore the data will be the same in both applications.

    >
    > Of course, you will have some limitations on your data caching options, but those will be a problem no matter what persistence technology you use.

    Yes do different clusters(unaware of each other) but the same application is deployed and used on both the clusters. The Entity Beans will be using the same database ?

    I understand we can't use any caching but is there any other problems that could occur. If we were to use CMP we can't dictate when the data gets stored would this cause any inavlid data in the database ? would BMP work ?

    And what Isolation levels on the transactions should be used (Read_Commited, etc)...Thanks
  4. Hi Max,

    Even with your current environment (i.e. two separate clusters running the same application) you _can_ cache data. Using Coherence's clustered caching you can cache data across the two clusters as well as keep it in synch with the DB. Coherence supports Transactional Caching (through a J2CA Adapter) allowing the developer to access the cache within a transactional context.

    Later,
    Rob Misek
    Tangosol, Inc.
    Coherence: Cluster your Work. Work your Cluster.
    Coherence Forums.