Administer > Configure High Availability Mode

Configure High Availability Mode

The typical configuration for a high availability environment has two or more CMS UIs that connect to the same UCMDB server, or to the same load balancer that manages a cluster of UCMDB servers. The CMS UIs are configured to work behind a load balancer, meaning that the load balancer serves as the entry point to the CMS UIs. All the CMS UIs are active at any given time and can handle requests from users. Requests are distributed to the CMS UIs in the cluster by the load balancer, and are shared evenly among all the CMS UIs.

Note  

  • The load balancer used for high availability must have the ability to insert cookies and must be able to perform health checks ("keepalive").

  • These instructions are certified for the F5 BIG-IP version 10.x and F5 BIG-IP version 11.x load balancers.

    If you are using a different load balancer, the configuration should be performed by a network administrator who has broad knowledge about how to configure load balancers, and principles similar to those described here should be applied.

  • This procedure assumes that you already have at least one CMS UI installed and configured.

To set up a high availability environment:

  1. Create a cluster of CMS UIs.

    Install one or more additional CMS UIs to create a cluster. For information about installing the CMS UI, see Install.

    Note  

    • The machines used for all of the CMS UIs in the cluster should have similar hardware, including the same amount of memory, and should be running the same operating system.
    • CMS UIs in the cluster must use the same port number for HTTP, HTTPS, and so on. You cannot configure multiple CMS UIs for high availability if they are using different ports.

  2. Start the CMS UIs.

    1. If the first CMS UI is not started, start the process and wait for the startup process to complete.

    2. Start the additional CMS UIs.

  3. Configure the load balancer.

    The load balancer is used to balance the load that is sent to the CMS UIs in the cluster. Configure the load balancer as follows:

    1. On the load balancer, configure a Cluster VIP address that sends requests to the entire CMS UI cluster.
    2. Configure a pool of back-end servers that represents all CMS UIs in the cluster. This pool is monitored by a health monitor, which sends requests that can be processed by any CMS UI in the cluster.

    3. Configure the health monitor by providing a keepalive address. The health monitor checks for the keepalive page of each of the CMS UIs. For the Cluster VIP address, use the following URL for the keepalive address: /ucmdb-browser/public/ping.jsp.

      Possible responses from this URL are:

      Response HTTP Response Code
      Up 200 OK (desired response)
      Down 503 Service unavailable
    4. Connect the health monitor to the CMS UI pool configured in step b.

    5. Configure session affinity on the load balancer:

      • Configure the load balancer to insert cookies to the responses that are sent back to CMS UI clients (internet browsers).

      • Using the Insert method, add a persistence profile cookie for the VIP address.

      Note  

      • The cookie name and value are unimportant, as long as the load balancer knows how to maintain stickiness with the cookies that it sends out.
      • F5 BIG-IP version 10.x adds a session cookie only to the first response per connection to the server. Other load balancers you might use add a session cookie to each response.
    6. If the VIP is configured to accept secure connections and the load balancer forwards the requests to the CMS UIs over HTTP, you must configure redirect rewrites. In the load balancer software, enable the Redirect Rewrite select All option to configure the HTTP profile associated with the Cluster VIP.

Send Help Center feedback