So... This issue actually results from two breaking changes:

First, as was already mentioned, as part of HHH-11953 Closed the behavior when a cache is not found was changed to throw an exception by default. This is in the process of being corrected with a more conservative change as part of HHH-12649 Open .

Second, the name of some "global" caches (query results caches, etc.) was changed.
See for example [how the timestamp region was added before](https://github.com/yrodiere/hibernate-orm/blob/59c3baae3271247bed516fe50952b65be1a27e5b/hibernate-core/src/main/java/org/hibernate/internal/CacheImpl.java#L79) and [how its name was defined](https://github.com/yrodiere/hibernate-orm/blob/cc51bbc639137247281d68d2cd7090be5fa49247/hibernate-core/src/main/java/org/hibernate/cache/spi/UpdateTimestampsCache.java#L36), and compare it to [how the region is added now](https://github.com/yrodiere/hibernate-orm/blob/b8674563d27611c09ed50086b2b4534031f07204/hibernate-core/src/main/java/org/hibernate/cache/internal/EnabledCaching.java#L80).
Same for the query results region: [how the query results region was added before](https://github.com/yrodiere/hibernate-orm/blob/59c3baae3271247bed516fe50952b65be1a27e5b/hibernate-core/src/main/java/org/hibernate/internal/CacheImpl.java#L84), [how the region is added now](https://github.com/yrodiere/hibernate-orm/blob/b8674563d27611c09ed50086b2b4534031f07204/hibernate-core/src/main/java/org/hibernate/cache/internal/EnabledCaching.java#L89).

This second issue is annoying on multiple levels:

1. First it means we're breaking user configuration, since users may have configured a cache with the "old" name, and we're now ignoring this configuration.
2. Second it means we're now using the name of SPI classes for cache names which have to be manipulated by user; if we ever change those SPIs by renaming the classes for example, it could impact users.
3. Third, as we can see in this ticket, when combined with HHH-11953 Closed , this means users always have to explicitly configure these caches with weird, SPI names, whenever they use features that require these caches.

I think we should at the very least revert the name of those caches to use the previous names, which would solve 1 and 2. If we also fix HHH-11953 Closed , this should get rid of 3 too, and then we should be good.

In a future version we may choose to change the names of these caches, but it's clearly not something we should do lightly.