Will data be lost if the NAS system drive breaks? I reinstalled Feiniu, and 40,000 photos were not missing

At eleven o'clock

night, my wife's phone popped up with a "Backup failed" message. I opened the Feiniu app and spent a long time browsing. I pinged the NAS internal network address, but it didn't connect. That N100 had no screen, so I couldn't even see where it died.

clarify first: the system disk and data disk are separate entities

things I should have figured out on the day of installation, I only truly realized on the night the system drive crashed: NAS system and data are stored separately . When installing Feiniu, it will ask you to select a drive to use as the system drive, where the system, packages, and account information are all stored; Your movies, photos, and downloads are all stored in a storage pool made up of separate data disks. The water from the east and west wells is diverted twice, so it does not encroach upon the river water.

here's my setup: a 64GB Lao Kingston SSD as the system drive, taken from an obsolete computer years ago; A 4TB Cool Wolf is the main storage, with 46,000 photos and a 186GB phone photo backup library all on it; There's also an old SSD dedicated to download drives, all the QB work is handled by it.

So, "system drive failure" and "data lost" are two different things. What truly determines the fate of your data is which disk is faulty, and whether your hands are steady when reloading.

my family's time: the 64G system drive was gone and then disappeared

tell me the order of the checks that night, and you'll know them well." The app won't connect, so first check the router's backend device list—this N100 hasn't come for IP for several hours. Power out, wait half a minute, power on again, wait another twenty minutes, but it's still not in the device list. At this point, it's basically clear it's not a network issue, but the machine itself not starting.

the next day, I disassembled the computer, unplugged the 64GB system drive, and inserted it into the hard drive enclosure to connect the computer. CrystalDiskInfo pops up as soon as you open it, a warning pops up, SMART info can't be read across the entire page, and the health bar is simply empty. A single disk can't even hand over its own medical report; it's basically a controller or firmware issue, nothing to fix.

ironically, I stared at that 4T Cool Wolf SMART for a whole year, memorized the red and yellow card indicators by heart, wrote an article on reading methods, but never looked at the system disk. After all, it's called a "system" drive. Subconsciously, I treated it as part of the machine, forgetting that it's also a hard drive that ages quickly.

Full Step-by-Step of Heavy Installation Feiniu: Removing the Data Drive Saved Me

Before reinstalling

, the first thing I did wasn't install the boot drive, but power off and unplugged both data drives . This step is the most valuable experience in the entire process. Because the installation interface asks you to manually select the target drive. A row of drives is placed there, and if your hand shakes and you pick the wrong one, you click formatting. 46,000 photos are cremated on the spot—not even a god can save you. Physical disc removal is the dumbest yet safest foolproof feature—if the disc isn't in the machine, you have no choice.

the subsequent process is straightforward and straightforward. I wrote the USB drive into Feiniu's ISO with Rufus to make a boot drive, plugged it into the network cable to power on the monitor, and the brand new 128G SSD cost 89 yuan—make sure to install it. Installation takes about twenty minutes. This time, I'll turn on the system drive with it plugged in separately, confirm the system can enter, then turn it off and plug the two data drives back in.

Step 6 is a pitfall I had previously checked: the reinstalled new system does not automatically recognize the old storage pool ; you have to enter storage management and manually mount the original hard drive and storage pool back. All the data is on the drive, not a single layer is missing, but the system doesn't actively recognize it. If you don't click it, the file manager is completely empty. After uploading, I opened the photo catalog—46,000 photos, not a single one missing. That night, I finally slept soundly.

Reinstalling Isn't All Is Right: Five Things I Did Again

data is back, but don't think it's all right. The other side of the system and data separation is that the configurations on system drive are cremated along with the old disk. Which items were revived with the data, and which needed to be redone—I'll first cross-check this account for you. Survived: All files in /vol1, Docker containers mapped to the data disk, and downloads. Lost: user accounts, shared directory permissions, container definitions, scheduled tasks, network settings.

so from the night of the refurbishment to the next day, I did five things. First, create an account. My parents reopened two read-only accounts, reconfigured the shared directory permission matrix according to previous , checking who can only view photos, who can write download directories, and checking them one by one. Second, copy the parameters of the Docker container directly. I've written about the old three things—port mapping, directory mounting, and environment variables—before. Following the parameter copying routine pull them one by one, and the QB download directory points to the original location of the data disk. The record is complete.

the third item, re-reattach the album directory and re-run the index. It takes over a day to scan 46,000 images with the N100, and the CPU buzzes all day—there's no way to avoid this. Fourth, reconfigure the disk water level and SMART's inspection script, and connect the alarm channel back to Bark. Otherwise, I'm the last to know if it fails next time—last time I set up an alarm, I suffered losses, backup silence failed for three nights that time. Fifth, learn your lesson this time: export the system configuration as a bin file in System Settings, save one copy each to the data drive and the cloud drive. Next time you reinstall, you won't have to manually do these five things.

Synology's logic is different: the system is hidden in every disk

cousin's second-hand four-seat Synology car, the bad habits are completely different from mine. DSM's system partitions mirror each disk, so even if you remove one drive, the system can still work. This is where it is more robust than single-system disk design. When DSM really doesn't work, the reinstallation process is different: when you open the find.synology.com browser and search for the machine on the local network, the page will prompt "DSM can be installed." Choose a version that matches the original system reinstall it. The installer only rewrites the system partition, not the data partition.

really need to be wide-eyed at the installation option step. A pop-up will ask you to choose "Preserve system configuration and data files" or "Keep only files"—the former repairs the installation and retains all configuration data; The latter is equivalent to reinstalling, keeping the files but wiping out the configuration. If you choose the wrong step at this step, your account and package configuration will be explained. By the way, let's clarify two things: this is called a reload, which is different from the usual system update that pops up. An update is a live upgrade, while a reinstall is a system layer overhaul and reboot; The guide for Black Synosys is even more unconventional, but that's a matter for another article.

suffer a setback: the system is a scam, and you have to keep an eye on it

After this incident, I put all three SMART drives at home for inspection, and the system drives are not exempt. Old SSDs have a hidden risk when used as system drives: they often fail without any warning before they fail, and mechanical drives at least give abnormal noises or red card indicators to slow things off, so the SSD controller can just leave at any moment. So that 64GB Kingston has served five years without a checkup, which is just my luck.

two things have now become routines. System configuration exports the bin once a month, stored together with the data disk backup; The foundation for important data is still the old 3-2-1 rule: two local copies plus one off-site account, backup accounts have already calculated . Because "reinstalling to protect data" has a premise you must remember—it only covers cases where the system drive fails. If the damage is the data disk itself, or if you accidentally formatted the data disk during installation, how much you can recover depends on how reliable the backup is. If it really comes to that, you'll have to try accidental deletion recovery and try your luck.

to wrap things up in order: If you can't connect, first distinguish whether it's a network or a machine. For machine issues, disassemble and connect to the computer to test SMART. Before reinstalling, always remove the data disk. After installation, manually mount the storage pool. Finally, reassign the account container alert, and then import a configuration bin. The first time I went through the full process, it took me half a night and half a day. Now I have configuration backups, so doing it again might take about an hour.