雖然企業版HDD o既MTBF比普通版高唔少, 但bad sector remapping o既持性唔一樣.
所以用o係single disk/RAID 0 environment o既話有必要考慮埋:
From Synology's wiki:
http://www.synology.com/wiki/ind ... a_Desktop-Class_HDD
Most or all HDDs which are shipped today have defective sectors, or will begin to have defective sectors as the HDD ages. HDDs today have a built-in ability of repairing defective sectors with a backup supply of sectors. This repair process is seamless to the user, and is not required for the user to perform sector remapping manually.
However, given the two aforementioned environments, sector remapping times vary greatly between the two. With a Desktop-class drive, the HDD will attempt to retrieve the data and remap the sector which is damaged, requiring as much time as several minutes. During this sector remapping process, a HDD can time out and become unresponsive to the HDD controller. Because of this, a HDD can be ejected from the controller, as the Controller believes that the HDD is defective as it is unresponsive.
This effort of using several minutes in attempting to retrieve data from a defective sector is essential for single-drive usage, as there is not another copy of the data available. However this behavior is not beneficial in RAID environments, as a drive being ejected from the RAID Controller will result in the Volume becoming degraded, due to a few bad sectors.
To resolve this, sector remapping for Enterprise-class HDD is handled differently. On an Enterprise-class HDD, the HDD will attempt to retrieve the data from the affected sector, however if it cannot be retrieve under a few seconds, the HDD will mark the sector as defective and consider the data lost. The HDD maintains a lower time interval to perform sector remapping thus to avoid timing out or being ejected from the RAID Controller. The HDD will then ask the controller to rebuild the missing data from the other Redundant drives (RAID 1, 5, 6 for example); the rebuilt data from the other drives will then be placed on a backup sector of the HDD. If an Enterprise-Class HDD is used as a single entity, or used in a RAID-0 environment, then the data on the sector which was damaged is considered lost as there are not other redundant mirrors to rebuild the affected data.
In summary, it is not recommended to use Enterprise-class HDDs as a single entity or in a non-redundant RAID environment (JBOD, RAID-0 for example)... |