Announcement Announcement Module
No announcement yet.
tc Server : group vs cluster Page Title Module
Move Remove Collapse
Conversation Detail Module
  • Filter
  • Time
  • Show
Clear All
new posts

  • tc Server : group vs cluster

    Hello all,

    I'm new to this and bare with me if this sounds stupid

    I setup two tc Server instances in one machine and I grouped them using HQ Hypric, setup load balancing using ERS httpserver..deploy applications and all works fine..

    now, the question is what is the differences between grouping and clustering?

    as far as i can see the load balancing works..and fail over maybe? I tried take one instance down and the httpserver will auto route the request to another instance in the group.

    Is this the correct way / best practice to do the whole load balance, HA..etc

    Thanks all!

  • #2

    Myself kanika i will also want to known what is the difference b/w group &,also reply me as soon as possible.


    • #3
      Grouping in Hyperic is from an administration standpoint. You can group any resource of the same type together (whether they are clustered or not).

      Hyperic grouping gives you aggregate metrics, and it gives you some group command functionality. For tcServer groups this includes group start and stop and deploy application war file to groups.

      Setting up a cluster in tcServer by using the cluster template when creating instances (or by adding the clustering config in server.xml) turns on session replication between nodes. You will need to have a load balancer with sticky sessions, you will need to implement java.serializable for all session objects you want replicated, and so forth. For more info on tcServer clustering, see the Tomcat doc on clustering. Clustering in tcServer is similar to clustering in Tomcat 6.

      Note: You can have load balancing and fail-over with tcServer without turning on session replication. Depending on one's definition this is also a cluster. In fact it is far more common to have load balancing and fail-over in place for scale up and high-availability (without session replication). Turning on session replication should only be done when absolutely required.

      Hope that helps,