1. Support the ability to re-open the backing strategy without re-opening the Journal. This probably implies a re-open of the write cache service, which is fine.
2. Subsequently, use this capability in abort handling if we had previously encountered an error. This will force the reload of all metadata from the disk.
3. Review other cases where the same code might be invoked. For example, after accepting the root blocks from the quorum leader in HA? (Though that code path seems to be fine).