RAID reliability: Toshiba explains whether it is possible to mix hard drives

I systems RAID (Redundant Array of Independent Drives) combine multiple disks into a single logical drive to improve data storage performance and reliability. Like any other electronic component, RAID configurations can be subject to failure, errors, and degradation. Optimization and maintenance of RAID systems are therefore of crucial importance to ensure reliable operation.

The hard disk used to compose the RAID use two possible schemes for managing the data stored in them. The acronyms 512n e 512e they refer to the block size of the sectors of a hard disk and represent two different approaches in the structure of the stored data.

Laboratory technicians Toshiba they wondered whether different types of hard drives could be mixed within the same RAID system and what effect this might have on performance and reliability.

The difference between 512n and 512e in the case of hard drives

Understanding the differences between the 512n and 512e scheme is important when designing them RAID configurations or select disks for specific storage needs. 512n hard drives represent the traditional approach, while 512e drives offer a solution that balances compatibility with the latest technologies and support for legacy applications.

On a hard drive with block size native 512n, data is written and read in 512-byte sectors. Drives of this type are often referred to as “native” or “non-emulated” disks.

Conversely, in a 512e hard disk, the drive simulates the structure of a disk with a block size of 512 bytes, but internally uses blocks of 4,096 bytes (4k). This emulation was introduced to improve compatibility with applications and systems designed to work with 512-byte disks, while maintaining larger internal block sizes.

In RAID systems, compatibility between disks with different block sizes is an important consideration, as it affects the ability to replace, combine, or expand the disk array.

What you need to know when setting up or modifying a RAID system

The technicians of Toshiba HDD Laboratory recently evaluated the flexibility of 512n and 512e block sizes in RAID systems. During the tests, the experts tried to answer several questions:

  • Are 512n and 512e block sizes fully interchangeable in RAID systems?
  • In the absence of a hard disk identical to the one to be replaced, will a disk with a larger capacity still perform the task?
  • What happens if it is not possible to find a hard disk identical to the one in use? Is it possible to use a similar disc from another manufacturer?

Carrying out the test

Toshiba experts have used native and emulated hard drives in NAS and Enterprise Storage: the popular Toshiba NAS HDD N300 as the representative 4TB model with 512n block size and the Enterprise model of the series MG-Series MG08ADA400E with block size 512e.

Hard disk Toshiba N300 per NAS

Per simulate a failure, technicians simply unplugged the energized units while they were running (hot removal). After waiting 10 minutes, they proceeded to test the performance and functionality of the RAID array. The next phase took place with the subsequent insertion, always under power, of the replacement hard disks. Obviously, the system has been given all the time it needs to correct rebuilding the RAID array.

The ratings were based on performance highlighted with two models of NAS with 4 rather common compartments: QNAP TS-464-4G e Synology NAS DS-420+. Several were also used controller RAID popular, with hard drives installed in the front bay of the server (Microchip Adaptec SmartRAID 3204, Microchip Adaptec SmartRAID 3154, Broadcom MegaRAID SAS 9460 and its successor MegaRAID SAS 9560).

Synology DS420+ NAS

It is possible to mix 512n and 512e hard drives in RAID configurations

The results of the tests conducted by Toshiba highlight that, for all configurations, i failed 512n disks they can be replaced with 512e models without limitations. The opposite is equally true: replace a failed disk 512e with a 512n drive it works in all cases. Thus, mixing 512n and 512e is possible for all RAID configurations, although the reconstruction strategies differ between cases.

A failed disk can be replaced with a model of superior capacity, but the additional capacity provided by the replacement drive remains unused. These tests also indicated that hard drives with the same interface (SATA), same block size (512 bytes, regardless of whether it is emulated or native), and equal or greater capacity can be mixed, regardless of manufacturer.

Furthermore, in general, the reconstruction time of RAID with a legacy, therefore rather dated, controller is not that different from that found with more updated RAID controllers. Toshiba suggests that RAID reconstruction performance is therefore predominantly dependent on disk speedrather than the performance of the RAID controller itself.

Adaptec SmartRAID 3204

Toshiba also used RAID controllers such as the Microchip Adaptec SmartRAID 3204 as part of the test platform.

The other relevant aspects that Toshiba highlights following the test

The possibility of mix discs native and emulated in virtually any RAID configuration is good news for users and operators. However, there are other important aspects to consider. First of all, it should be remembered that the RAID technology it is not a backup: it allows you to continue working in the event of failure of the storage medium.

If the RAID controller/box fails or other storage system problems occur, you risk failure data loss. Furthermore, a failure on a single hard disk cannot be identified simply by observing the operation and performance expressed by the RAID system. It is therefore important to regularly check the indicator LEDs on RAID boxes and NAS systems. For the latter it is also possible to refer to administration panel Web.

Equipping yourself with replacement discs in advance is always advisable

A replacement disk should be purchased in advance: ideally, it should be the same model as the original RAID set but, as Toshiba’s tests show, there is a lot of flexibility in the end. The important thing is that it uses the same technology, has the same or higher capacity than the drive being replaced, and does not have a native block size of 4k.

After replacing the failed disk, the system should be checked to see if automatic rebuilding has started. Otherwise, you must start it manually by accessing the appropriate menu. While you can continue working, it is best to leave the RAID system idle as this will shorten the rebuild time.

Failed disks also shouldn’t never be reinstated in the RAID system. If a disk has failed once, it will certainly create new problems. And even if everything seems to be fine after reinsertion, many RAID systems get confused if they recognize a disk that was part of a previous RAID configuration (due to the metadata present). The results can be catastrophic.

To learn more about the results of the research conducted by Toshiba, we suggest referring to the new whitepaper, which can be downloaded for free by clicking here. In the article dedicated to how to choose the best QNAP NAS, we also focused on the differences between the various RAID configurations.


Please enter your comment!
Please enter your name here