Pinned data can occur when issues arise regarding the Drive Enclosure storage array. In such a case, data that is to be written to that array remains in the flash-backed memory of the Controller.
Each LUN maintains a sequential record of committed transactions (a set of modified blocks) that are guaranteed to be written to the underlying drive group storage. The Oracle FS System maintains one cache for each LUN. In the background, the system asynchronously flushes each cache to permanent storage. These records are managed within what are called write cache.
The system maintains two copies of the cache in flash‑backed memory. The primary copy exists on the same Controller node on which the LUN resides. The secondary copy (a mirror) exists in the flash‑backed memory in the partner node to allow recovery in the event of a failure of the home Controller.
An administrator-initiated shutdown request fails if any user data is still cached and has not yet been written to physical storage. Pinned data causes the shutdown to fail. If the Controllers cannot communicate with the Drive Enclosures to flush the cached data, the Oracle FS System retains, or pins, the data in cache.
If you receive a system alert about pinned data when you initiate a shutdown request, navigate to to review the details about the Drive Enclosures. Resolve any hardware issues that might exist. Hardware issues can prevent communication between Controllers and Drive Enclosures and can prevent the system from flushing the cached data to storage.