The explosive growth of location-detection devices, wireless communications, and mobile databases has resulted in the realization of location-based services as commercial products and research prototypes. Unfortunately, current location-based applications (e.g., store finders) are rigid as they are completely isolated from various concepts of user "preferences" and/or "context". Such rigidness results in nonsuitable services (e.g., a vegetarian user may get a restaurant with non-vegetarian menu). In this paper, we introduce the system architecture of a Context and Preference-Aware Location-based Database Server (CareDB, for short), currently under development at University of Minnesota, that delivers personalized services to its customers based on the surrounding context. CareDB goes beyond the traditional scheme of "one size fits all" of existing location-aware database systems. Instead, CareDB tailors its functionalities and services based on the preference and context of each customer. Examples of services provided by CareDB include a restaurant finder application in which CareDB does not base its choice of restaurants solely on the user location. Instead, CareDB will base its choice on both the user location and surrounding context (e.g., user dietary restriction, user preferences, and road traffic conditions). Within the framework of CareDB, we discuss research challenges and directions towards an efficient and practical realization of context-aware location-based query processing. Namely, we discuss the challenges for designing user profiles, multi-objective query processing, context-aware query optimizers, context-aware query operators, and continuous queries.