In this paper, a full-stack perspective that involves content coding, caching and recommendation is presented for wireless edge caching networks. Therein, content coding and recommendations are utilized to ensure the reliability of the network and to boost the gain of edge caching, respectively. Specifically, with a well-designed coding pattern, the concurrently occurred data corruptions of multiple access points (APs) can be successfully recovered. Whilst, the successful content retrieval is warranted with highly selective retrieval sets. Further, by offering personalized recommendations, the cache hitting at network edges can be promoted. On these grounds, a cache hit ratio maximization problem for reliability-aware wireless edge caching system is formulated, considering the successful content retrieval and repair, total repair cost requirements, cache capacity budget of each AP, and recommendation quality and quantity per user. To solve this intractable problem, a framework with two steps is developed, which leads to construction of a suitable coding schema with the minimum repair cost per content to satisfy the retrieval and repair criteria, thus achieving a stabilized optimal solution regarding the joint cache placement and recommendation decisions from a game-theoretical viewpoint. Finally, extensive numerical simulations are performed to confirm the superior performance of our devised algorithms compared with benchmark baselines.