You can optimize a site’s performance after a switch deployment by enabling selective cache invalidation on target repositories. A repository that is thus configured invalidates its item caches selectively during deployment, rather than invalidating the contents of those caches entirely.

If enabled:

Only those cached items that change as a result of the deployment are invalidated. All other cached items remain unchanged. Because selective invalidation increases deployment overhead, you might want to configure a threshold for the number of invalidated items. On exceeding that threshold—the sum of all changed items in a deployment—the selective invalidation process is skipped and the deployment invalidates all cached items in the target repositories.

If not enabled:

Item caches of all deployed repositories are invalidated. This can result in slow response time to initial requests, as fresh data is obtained directly from the database.


Two general constraints apply to usage of selective cache invalidation:

Configuration Steps

You configure selective cache invalidation on the production server and the asset management server, as follows:

Excluding Repositories and Item Descriptors

A production site publishing agent can be configured to exclude specific repositories and item descriptors from selective cache invalidation—in other words, require that item caches be fully invalidated on each deployment. To do so, set the property fullInvalidationRepositoryPaths in this component:


You set this property as follows:

   repository-path[=item-descriptor[;item-descriptor]...] \

If no item descriptors are supplied, all item caches in the repository are invalidated. The property can specify multiple comma-delimited repositories, and each repository can specify multiple semi-colon-delimited item descriptors.

For example: