Restore fails while there is a background mutations happening in parallel.

1. Load items into the cluster (Spring was used in the testing)
2. Once the load is complete, start a load on the data loaded (50% read, 48% updates, 1% creates and 1% deletes).
3. Take backup of the cluster. (repeat this step multiple number of times)
4. Flush the bucket.
5. Restore the cluster with the backed up data. (dont mention the start or end point, let restore happen on all the backups.)