Failover and data recovery in caches

If or when a cache server becomes unavailable, the web application nodes continue indefinitely to attempt to communicate with the cache server until it is brought back online.

A web application node determines that a cache server is unavailable after 20 consecutive failed requests to the cache server. The web app node then waits for two seconds before attempting to communicate again with that cache server. This process continues indefinitely until the failed cache server is brought back online. In the meantime, the web app nodes automatically redirect to the next available cache servers.

If the cache server is unavailable at the startup of a web application node, the web node uses local caching. The web app node continues to try to communicate with the external cache server, and does so indefinitely, until the external cache server is brought back online.