There are currently a lot of heap allocations when deserializing vBucket maps due to list resizing. Given that we have advance knowledge about the structure of the maps we can optimize this.