Synology Hybrid RAID: Storage layout and specifics of data recovery
The release of less expensive NAS models along with the introduction of SHR in 2010 has contributed to the increasing popularity of Synology among home users and small businesses who became able to effectively store and share digital content without the need for thorough planning, performing any calculations or delving into complicated RAID concepts like mirroring, striping or parity. However, in spite of their robustness and convenience, such NAS appliences also develop problems that result in data loss and to successfully restore the missing files, one should have a grasp of data organization peculiarities typical of Synology RAID which directly impact the process of recovery from network-attached storages based on this technology.
What is SHR?
Synology Hybrid RAID or simply SHR is an alternative vendor-specific method of redundancy designed to combine a number of Synology NAS drives into a single logical unit with the aim to protect the device against data loss caused by disk failure while overcoming considerable limitations of classical RAID, among which is total lack of flexibility and rather complicated setup.
SHR was released in March 2010 together with DiskStation Manager 2.3 and is now supported in most Synology NAS models. Based on the combination of Linux mdadm and LVM, this data allocation scheme offers for more automated storage management, improved scalability and allows mixing drives of different capacities in one NAS box while practically all available space can be effectively used for keeping data.
Unlike RAID, SHR doesn’t have so many variations. Two major options are available: SHR and SHR 2. SHR protects the array from a single drive failure while SHR 2 – from two simultaneous disk failures.
Basically, we can draw the following parallel between SHR and traditional RAID levels:
SHR consisting of 2 drives corresponds to RAID 1;
SHR composed of 3 or more drives is similar to RAID 5;
SHR 2 can be compared to RAID 6.
Hint: All the basic RAID concepts are described in the peculiarities of data organization on RAID.
However, it should be mentioned that such an analogy is only valid from the perspective of storage use while internal data organization principles are not always exactly alike.
Each physical disk constituting Synology RAID is divided into several smaller allocation units based on the size of the smallest drive in the array and then software RAID can be automatically created across these smaller sections instead of necessarily using the entire disks.
That being so, if disks of the same capacity are used, the result won’t actually differ from conventional RAID (levels 1, 5 or 6, depending on the number of component disks and chosen level of redundancy). However, if the disk sizes do not exactly match, a sort of mixture of RAIDs is created in each case which enables utilization of the leftover space.
For instance, a four-bay unit has four drives: two are of 2 TB and the rest two are of 6 TB. In a classical RAID 5 all of the drives will be seen as ones of the same size which equals to the capacity of the smallest disk (4 x 2 TB = 8 TB) while the remaining storage space (2 x 4 TB = 8 TB) will be simply wasted. We also have to subtract one drive (2 TB) which will be used for storing parity and thus we will get 6 TB of usable storage space out of 16 TB.
In SHR, RAID 5 will be first established while the remaining “disk tails” (2 x 4 TB) will be organized into RAID 1 and both RAIDs will be then spanned into single storage – JBOD. In total, SHR-based NAS will provide 10 TB of usable storage space with the protection from the failure of one of the disks in the array and consist of RAID 5 and RAID 1 merged into one JBOD.
Advantages and shortcomings
Synology Hybrid RAID can boast a number of the key strengths in comparison to classical RAID which is also available as a configuration option for most Synology NAS models:
hard drives of different manufacturers, speed and capacities can be mixed in one SHR;
SHR made up of dissimilar capacity drives provides a much bigger amount of usable storage space when compared to RAID – only one largest drive is “lost” for redundancy;
even so, Synology RAID guarantees protection against one failed drive and the subsequent data loss (with SHR 2 – even two disk failures);
new drives can be added to SHR anytime – no preliminary calculations are required;
However, SHR can be characterized by a slightly lower speed. Also, a rather complicated structure may become a hindrance to retrieving critical files in case of data loss.
When is recovery required?
Being predominantly based on the RAID technology, Synology Hybrid RAID inherited its vulnerable points and this may lead to inaccessibility of important files:
an inaccurate configuration may be caused by the loss of electric power or improper rebuild;
file system corruption can be a result of software crashes, malware infections, power outages, improper shutdowns, etc.;
firmware or operating system errors may make certain partitions or the whole storage inaccessible;
user errors, such as mistaken deletion, formatting, reinstallation as well as careless configurations may also bring about a variety of data loss issues;
physical failure or damage to NAS box or its disks may result in total data loss, as such systems are designed to tolerate a failure of up two drives (depending on the configuration). The main signs include the inability of the system to recognize disks, unavailability of the user interface, error messages, unusual color of light bulbs on a NAX box, beeping, buzzing or ticking noises, etc.
Data recovery process: recommendations and precautions
What is needed to perform the task?
Data recovery from SHR-based NAS is a complicated procedure which requires advanced software tools present only in the professional and RAID editions of UFS Explorer. For this process you will also need:
a personal computer running Windows, Linux or macOS, which will be used for installation and running the software;
enough free space on the PC used for recovery or any external storage medium or network storage;
SATA cables and free ports on the motherboard to connect NAS drives to a recovery PC. USB-to-SATA can be used as well.
Hint: Please consult the instructions to safely remove the disks from your NAS and connect the drives to the computer for data recovery.
How should one assemble the array?
Metadata on the Synology RAID drives is automatically recognized by the software and the array will be reconstructed in the virtual mode.
Yet, the process of data recovery can be hindered by the loss or corruption of SHR metadata. In this case, the software won’t be able to reconstruct the configuration automatically and the user will have to define the array manually. For most Synology NAS only “data partitions” – the largest partitions of each physical disk constitute RAID and should be added as its components.
If disks of the same size are employed in SHR, the data from the system can be recovered like from any traditional RAID. If two or more drives are of bigger capacity, their tails are to be assembled into another RAID and then spanned with the main RAID. In fact, two arrays get defined separately and added as constituents of a JBOD. The main stumbling block here is to find the right tail components and correctly span them with the main RAID.
What safety measures are to be taken?
Attempts to rebuild the system after data loss should be avoided as they can compromise the chances for successful recovery.
A good piece of advice is to label the drives according to their order before taking out them from the NAS box. In a vertical layout disks are numbered from the top to bottom and from left to right in the horizontal one.
In case of hard disk drive failure, the failed drive(s) can be determined with the help of the Storage Manager utility from DiskStation Manager. In the left-hand sidebar choose HDD/SDD which will show the list of drives and their current status. During data recovery, the failed disk(s) have to be substituted by virtual placeholder(s) so that the software could reconstruct them using redundancy.
Last update: August 18, 2022