Due to the fact that we cannot query in-use images, some time ago we've started storing the total disk image allocated size within a JSON file. At the same time, we're caching those values within the service. As we're aiming to support A-A deployments, also considering the risk of having this file corrupted, we're now dropping it. Instead of using a separate DB, we're just going to rely on the Cinder DB when retrieving the allocated space for a specific share. One reason why we did not go with this approach from the beginning is the fact that we needed support for reporting each share as a pool, which is accomplished by the other patches from this chain. Change-Id: I2085c040d7140f8ba685f6a36c4d3b5b30b14103
247 B
247 B