Take the Read error rate – 200(51) -- as an example, the value 200 is the normalized value and the value 50 in the parentheses is the threshold value. All these values are defined by the disk vendors and variant between different vendors. Even different models come form the same vendor may have different definition of S.M.A.R.T. The normalized value is just like a score and higher score is better. As to the threshold value is the bottom line of the attribute. When a normalized value is under the threshold value means this disk may be failed at any time.

Again, the S.M.A.R.T. values show on the Web UI is provided by the disks themselves, Thecus does not judge the disk health by S.M.A.R.T value. It is provided as a reference to users by the disk vendors; we just read and show them on the Web UI and leave the judgment of S.M.A.R.T. status to the users. The way we diagnose the disk health is according to the status when we have the I/O commands with it (e.g. I/O time out, I/O retry, read error…).