Configure load-balancing and/or high-availability

To use the Web Services Proxy in a highly-available (HA) configuration, you can configure load balancing. In an HA configuration, typically either a single node receives all requests while the others are on stand-by, or requests are load-balanced across all nodes.

About this task

The Web Services Proxy can exist in a highly-available (HA) environment, with most APIs operating correctly regardless of the recipient of the request. Metadata tags and folders are two exceptions, because tags and folders are stored in a local database and are not shared between Web Services Proxy instances.

However, there are some known timing issues that occur in a small percentage of requests. Specifically, one instance of the proxy can have newer data faster than a second instance for a small window. The Web Services Proxy includes a special configuration that removes this timing issue. This option is not enabled by default, because it increases the amount of time it takes to service requests (for data consistency). To enable this option, you must add a property to an .INI file (for Windows) or an .SH file (for Linux).

Steps

  1. Do one of the following:
    • Windows: Open the appserver64.ini file, and then add the Dload-balance.enabled=true property.

      For example: vmarg.7=-Dload-balance.enabled=true

    • Linux: Open the webserver.sh file, and then add the Dload-balance.enabled=true property.

      For example: DEBUG_START_OPTIONS="-Dload-balance.enabled=true"

  2. Save your changes.
  3. Restart the Webserver service so the change takes effect.