Tableau Server on Linux Help

The Cache Server provides a shared external query cache. It’s a cache of key/value pairs which hold information from previous queries to speed up future requests. To make Cache Server highly available, configure one or more Cache Server processes on multiple nodes of the cluster.

What happens when Cache Server process goes down? The consequences are relatively mild. Tableau Server will still work but actions may take longer as they do not have cached results available. As the queries rerun, the restarted Cache Server gets repopulated, eventually speeding things up for end users. In effect, Cache Server does not have an availability impact; however, it does have an impact on various end user performance scenarios.

To reduce impact on user performance, run multiple processes of this type across the cluster. We recommend you install no more than a maximum of six cache server instances, limiting each node to no more than two. Our testing indicates that installing more than a total of six cache server instances in a Tableau Server installation provides no improvement and could have a negative impact on performance.

A failed Cache Server process is automatically restarted; as long as the computer itself is otherwise healthy, the Cache Server process will relaunch.