Converting Two Standalone OO Central Installations into a Cluster

Hello,

We currently have two standalone OO Central 24.1 installations (Central1 and Central2) running on RHEL 8.9, but we are considering transitioning to a clustered setup. We've reviewed the official documentation but still have some questions:

  1. Is there a way to convert Central2 into a cluster node, or is a reinstallation required?
  2. While a Load Balancer is not mandatory, if we don’t use one and direct all requests to either Central1 or Central2, will both workers still share the workload?
  3. How will this change affect our Node License?

Thanks in advance for your help!

  • Verified Answer

    +1

    Hi,

    While technically it would be possible to convert one of the standalone centrals to a cluster node for the other node there are so many error points that can cause issues that a reinstall would be quicker. Keep in mind that even if you try to convert a central to become a cluster node it will not also migrate the database history to the cluster and as such the gain of conversion over reinstall would be 0.

    In a clustered  environment the workload will get distributed reasonably evenly between the worker regardless of utilization of a load balancer. The database holds the information about worker availability and queue of work to be done so either central could assign work even to a worker that is not connected directly to it or does not belong to it.  Without loadbalancer, however in case one of the centrals goes down access to that side of the infrastructures of workers will be lost and all RAS tied to that node will become offline.

    For Licensing issues this is a bit more complicated to predict, however having someone from OpenText  (not sure if this is handled by slaes, support or which team) do an audit of the license usage on both centrals as well as a check for overlap in nodes will better estimate what the new license for the unified environment would look like.

    Hope this helps,

    Vlad

  • 0 in reply to 

    Thanks a lot for the answers, Vlad!