Rust is a young programming language with safety memory features by its ownership rules. However, both human-being and automated code generators easily violate ownership rules, and these violations can result in compilation errors. Although rustc, the official Rust compiler, offers some repair hints in its compilation error messages, on a set of 160 erroneous Rust programs, our experiment shows that the repair hint only fixes 45 (28.1%) programs. To enhance the usability of Rust, we present rust-lancet, the first template-based automated Rust ownership violation repair tool with behavior preservation. We introduce two novel components: (1) an iterative template based repair framework, (2) a mental semantic to model developer intention for the erroneous program, and the corresponding symbolic executor to preserve behavior between the pre-patched and patched program. The experiment was conducted on our dataset, 111 bugs from rustc test suits, and 42 additional bugs from StackOverflow. Our results show that Rust-lancet outperforms rustc in success repair rate by 37.2%, 26.79% on gpt-3.5-turbo, and 11.76% on gpt-4. We show that rust-lancet achieves the best performance.