The concept of cyber-manufacturing systems (CMS) is at the nexus of manufacturing advances by an interdisciplinary effort from engineering, computer science, and information science fields. Future manufacturing will increasingly be characterized by complex, networked cyber-physical systems that may be instantiated in one physical location but distributed across many. However, this vision of an interconnected manufacturing environment ushers in the challenge of new security threats to production systems that still contain traditional closed legacy components. How to address the system failures caused by cyber-attacks remains one of the main success factors in the widespread adoption of CMS. The aim of this research is to investigate the resilience of such a system in the face of cyber-attacks. In general, "resilience" refers to the ability of an entity to withstand unforeseen environmental perturbations or disturbances while continuing performing its intended functions successfully. A resilient CMS is a manufacturing system capable of detecting, withstanding, and recovering from failures while still fulfilling its expected levels of service with acceptable levels of security, integrity, and profitability. Particularly, the resilience from failures caused by cyber-attacks is the focus of the research. A resiliency measure, evaluation of system impacts, and selection algorithm of resilient mechanisms have been explored.