Affects Version/s: None
Fix Version/s: 2.0-beta1
Number of attachments :
Try the following:
- create a store connection that points to a certain postgis database
- try to create a layer from it, see the listing of layers available (stop here, don't actually create it)
- go back and edit the store by chaning the database it points to (another postgis database)
- try to create again a layer out of the store, you'll be greeted by the same list as before. The datastore seems to not have been reloaded in the middle
I guess the ResourcePool should listen to catalog object changes and purge cached (or vice versa).
Now the DataStoreConfiguration page is actually calling resourcePool.clear(...) manually, but that does not seem to work fine, most likely because the pool expects an unwrpped StoreInfo whilst the page is having a proxy instead.
Anyways, it seems to me it's the job of some catalog event handler to purge the pool cache, not the UI, otherwise we'll end up replicating the same logic in the RESTConfig and whatever other subsystem may need to play with the configuration (e.g, and embedded ingestion engine)