What are the differences between Synology SHR and RAID1? I calculated the 4T and 8T separately, but choosing the wrong one wasted 4T
On the first night I
installing Synology, I got stuck: at the stage of building the storage pool, there was a row of options in the type bar—Basic, JBOD, RAID 1, and below it was an SHR. I had two disks of different sizes: one 4T Cool Wolf and one 8T Cool Wolf. If you pick the wrong one, it's a waste; if you pick the right one, you don't know if you're right. This article will break down the accounts of SHR and RAID 1 and clarify them. If your disk size is not the same, please click next after reading.
What exactly is
SHR? Don't be scared by the name just yet
SHR stands for Synology Hybrid RAID, Synology's hybrid RAID. Synology's official knowledge base clearly states its positioning: this is a RAID management system built on Linux, simpler than traditional RAID, suitable for those new to RAID; If you prefer manual RAID management, go for standard RAID.
its underlying layer isn't any black technology, just Linux-standard mdadm. Synology has layered an automatic separation logic on the outside: it automatically splits into several layers based on the drive capacity you insert, small drives pair up with small disks, and the rest of the big disk's life forms another layer. You don't have to calculate which disk to pair with which one to pair yourself; it does it for you.
By the way, when choosing a type, you might also encounter the SHR-2, which is its second-generation form, starting with four discs, allowing two to fail at once, with tolerance comparable to standard RAID 6. At home, with four trays and a sink filled with photos and other things that are lost and gone, spending a few extra trays on the SHR-2 is worth it.
So don't think SHR is a "more advanced RAID"—it's actually "RAID that automatically settles accounts for you." Whether you save on your accounts depends on whether the tray in your hand is the same size—this leads to the most crucial part below.
two disks of the same size, SHR and RAID 1 are no different
to clarify the most common situation: two 4TB drives, select SHR, and the output is 4TB images plus 4TB images, with 4TB available space, and the data on a broken drive is still there. It's almost identical to RAID 1.
online debates about SHR are mostly because users of the same capacity are worrying about others—to them, these two are really the same thing. So for two drives of the same capacity, just blindly choose SHR. The result is the same, and upgrading to larger drives one by one in the future is more flexible.
drives are different sizes, so the difference comes in: 4T + 8T billing
my machine is 4T plus 8T. At that time, I used the RAID calculator from Synology's official website to count both, and the results were shockingly poor:
| Disk Combinations | RAID 1 Usable | SHR Usable Parts of the | SHR with No Redundant Parts |
| 4T+4T | 4T | 4T | No |
| 4T+8T | 4T | 8T | 4T (the latter half of the 8T drive) |
| 4T+4T+8T | 8T (RAID 5 caliber) | 12T | 4T |
RAID 1's rule is to calculate the minimum amount: only the first 4T of the 8T slot is mirrored, and the last 4T is left idle. You spend 8T of money and use 4T of space.
The
SHR cutout is: the first layer mirrors the 4T drive and the first 4T of the 8T drive, while the second layer lets the remaining 4T of the 8T drive work independently. Available space is 4+4=8T, not a single cent wasted.
also explains the third row of three drives in the table: at 4T+4T+8T, SHR takes 4T from each of the three disks to form a checksum strip layer (targeting RAID 5; if one disk is broken, 8T can be used), and the remaining 4T of the 8T drive continues to run alone. So the available space is 8+4=12T, which is a full 4TB more than standard RAID 5 on paper, but the cost is the same: 4TB is unprotected.
note, the extra 4T from SHR is traded for redundancy. If the first layer fails a disk, the 4TB data is still there; The second layer is just an 8TB drive, all alone. If this drive breaks, the data in the latter half is instantly lost, leaving no chance to save it. Many people think that once you choose SHR, 'a broken drive is fine,' but when it comes to mixing capacity, this is only half true.
here's a harsh truth: DSM won't let you choose which files to put in the security layer and which to put in the bare layer; the storage pool is scheduled uniformly. So don't expect meticulous operations like "placing important mirror areas or movies in non-redundant areas." Data security relies on backups. I've mentioned many times in NAS how to back up data: RAID manages bad drives, no matter if you accidentally delete them.
want to expand, SHR's rule is to swap tokens one by one
I ended up aiming for expansion: 4T+8T lasted over a year, and I wanted to swap out that 4T and swap it for an 8T, making it 8T+8T. The official Synology process is as follows: in Storage Manager, check which drive has the smallest capacity, shut down the computer, remove the smallest one, plug in the large drive, power on, and the storage pool enters repair mode. Wait for it to finish; You have to repair one piece before replacing the next. After replacing everything, there's an extra unallocated space in the storage pool, and you have to manually click 'expand' to really increase the capacity.
I've been through all three of these pitfalls:
- can't be impatient and pull out two pieces at once. Like RAID 1, SHR only allows one piece to be broken at a time; pulling two pieces is equivalent to breaking the pool directly.
- during the repair period, the entire pool is in a state without redundancy. If one more disk is damaged, the entire pool will be wiped out. Back up important data before trading; don't rush through repairs.
- the capacity won't automatically increase after repair. The first time I replaced the drive and stared blankly at the capacity bar, I realized I had to manually expand the unallocated space in the storage pool. This step is also listed separately in the official documentation, so missing it means waiting for nothing.
during the repair period, the machine isn't unusable—albums and media can still be accessed as usual—but it noticeably slows down—the mechanical drive is completely rebuilt, and read/write queues are queued. Repair progress is checked by percentage in the storage admin; I watched it climb from 3% to 100% all night. At this point, don't manually run backup tasks. Fighting for IO from both ends slows everyone and can drag out the rebuild time.
repair time should be mentally expected. Mechanical drives have a sustained write speed of over 100MB/s, and each 8TB drive takes at least ten hours to rebuild. I started repairing before bed that night and only finished running in the morning. I wrote in more detail about the specific process for replacing the drive and how to handle the old drivesNAS in on how to migrate data from a new hard drive. Before plugging in a new drive, it's recommended to go through the section of the new hard drive full inspection first, and don't swap faulty drives into the pool.
The two costs of
SHR should be carefully considered before choosing
first cost: not many people recognize this format. Standard RAID 1 can be directly recognized on Linux using mdadm. If the machine ever shuts down, you can unplug the drive and plug it into another device, and there's still hope for recovery. SHR is Synology's proprietary scheduling method; if you want to read it directly on another NAS or a regular computer, it's basically impossible. Black Synology is useful, but if something goes wrong, all the investigation materials are written in the DSM environment, doubling the cost of self-study.
second cost: changing environments is a hassle. Synology itself also recommends using standard RAID for enterprise environments, which is familiar with operations and cross-platform compatibility. Of course, for family users, it's the opposite. SHR is originally positioned for beginners and families, so this cost doesn't really affect the device at home.
performance is not a concern. The term "hybrid" might make people think it consumes the machine, but SHR's read/write speeds are in the same tier as standard RAID for disks of the same capacity, with similar CPU usage. Tiering is set during pool setup, so running normally is free of extra cost. I ran intranet copy on a 4T+8T pool, and SMB stabilized around 110MB/s. Gigabit network cards topped out first, and couldn't be compared to RAID 1 pools.
my opinion: for home use, I bought gradually and got bigger with each purchase, planning to upgrade and upgrade each unit in the future. SHR is worry-free; OCD wants every piece of data to have a mirror or the drive already has the same capacity, so go straight to RAID 1 and sleep soundly.
took the right place and wrapped up the
| your situation | I recommend choosing SHR | two disks |
| the same size. The result is the same as RAID 1, allowing flexible switching in the future | ||
| drives are different sizes. If you don't want to waste a large drive, | choose SHR. The extra space without redundancy can hold renewable data | |
| disks are different sizes, but all the data is important | choose RAID 1. Or add a disk of the same capacity and reassemble it | |
| which may later be migrated to another system or use Synology | Choose standard RAID 1, which is highly versatile |
the order is as follows: first count how many disks you have and see if their capacity is the same; When they're the same age, just close your eyes SHR; If the size is different, calculate it according to the table above, and figure out what to put in the layer without redundancy; When building the pool, I also decided on the file system. I wrote separately on this part in NAS drive using ext4 or Btrfs. As for how many market slots to start with, you can look back at the article NAS buy two or four slots. In the end, it comes down to one thing: neither SHR nor RAID 1 is superior—only your drive setup isn't up to par. The big RAID model account was previously covered in NAS RAID selection , so reading both side by side is even more comprehensive.
