Ransomware has emerged as a major threat to users, resorting to file encryption and system locking, demanding a ransom to release their files. Current mitigation techniques are reactive, leading to the loss of files before ransomware is identified. We propose R-Sentry, a lightweight and real-time solution using honey files as deception, for the early detection of crypto ransomware. Honey files are disseminated throughout the filesystem, which if accessed, will alert the detection system leading to proactive termination of the ransomware process before encryption begins. Determining the optimal placement of honey files within a file system is challenging. R Sentry determines the optimal placement of honey files across various folders by analyzing the file traversal patterns of existing and future ransomware variants. Empirically we show that ransomware samples that follow several identified traversal patterns are immediately detected. In particular, for a scenario with 10 folders comprising 80 files, the number of honey files placed in folders was as less as 15.