What is S.M.A.R.T.?

S.M.A.R.T. (Self-Monitoring Analysis and Reporting Technology) is a technology, developed by major hard disk manufacturers, included in most modern hard disks. This technology monitors and logs many events while the disk is running. Analyzing the data allows us to know the drive's reliability and predict its failure. The properties that are logged (called attributes) are dependent upon the manufacturer and the drive. The S.M.A.R.T. specification does not require that the manufacturer include any specific attributes. In fact the latest version of the specification does not even require any attributes at all. The manufacturers can now simply report the status of the drive. However, most manufacturers continue to support attributes and they determine which attributes are suitable and important to the drive they are designing. The attributes are not the always the same and, sometimes, even their meaning can be different. Most S.M.A.R.T. data is undocumented and each manufacturer can use an attribute in a different manner. Some manufacturers store power on times as hours, while others use minutes or seconds. But these attributes all have one thing in common: they can be used to know how the disk is performing and predict failures.
When we talk about failures, we must differentiate between two types: predictable and non-predictable. Predictable failures are those that don't happen suddenly, they can be caused by failure in the disk mechanics or damaged disk surface. The problem gets worse, until the disk crashes. Non-predictable are those failures caused by sudden events, like a sudden power surge or an electronic failure. S.M.A.R.T. can only help to detect predictable failures.
The attributes available for the disks can change from disk to disk and not all attributes are available for a disk. The values for each attribute are, most of the time, normalized values, with a maximum of usually 100, 200 or 253. The table in HD HeartBeat shows five values for each attribute:
? Current - This is the value read for that attribute
? Worse - This is the worse value found for this attribute
? Threshold - This is the minimum value for the attribute. If the attribute reaches the threshold value, the disk is about to fail
? Data - This field stores extra data for the attribute. What's in it depends on the attribute and the manufacturer. For example, attribute 194 stores the temperature in degrees Celsius. For some manufacturers, it also stores the minimum and maximum working values. Attribute 9 stores the number of hours, minutes or seconds (depending on the manufacturer) the drive was powered on, and others, like 5 store the number of reallocated sectors
? Flags - This field shows the flags associated with each attribute. This is set by the manufacturer and may vary from drive to drive. The flags can be a combination of these letters:
? F - this is a pre-failure check: values close to threshold indicates imminent failure. When the threshold is reached, the disk is about to fail
? C - result of online collection
? P - this measures disk performance (lower values indicate low performance)
? E - measures an error rate
? V - measures an event count - Ex. Attribute 5 indicates how many sectors had errors and were relocated, in the Data field
? S - attribute is preservable and is restored in each S.M.A.R.T. test
Lower values indicate a decrease in the reliability of the disk.
When analyzing S.M.A.R.T. data, the most important thing is to compare the current values with the thresholds. If these values are below the threshold and the attribute is a pre-failure check, the disk is going to fail very soon. If the attribute is a performance value, lower values mean lower performance, and usually this is due to some damage, and the disk is also likely to fail. You must not be alarmed if some attribute falls a little, many times this is a transient problem and it can be corrected with normal operation. The worst value is the minimum value that attribute has reached, it can tell you the variation for the attributes on your disk.
Another parameter to be considered is disk temperature. If your disk temperature is very high (usually above 60?C - 140?F), you should consider using a fan to make the disks in your system cooler. High temperatures reduce disk life and increase the likelihood of damage.
If you have a new hard disk and the S.M.A.R.T. analysis shows a problem in your drive, you should consider returning it to the manufacturer so it can be replaced with a new one.