Written 5 Sep 2026. This is a how-to, not a product card.
See if the old disk is dying
SMART reports whether the drive still looks healthy. Install smartmontools, list the disk, run a quick health check before you trust photos on a Marketplace PC. You don't need a NAS.
Only checking how big the disk is? That quieter look is see how big the disk already is. Only checking free room? see if the computer still has room. Disk already full and copies fail? when the computer's disk is full. This note is different: the drive's own health report, not how full it is.
Skip until SSH works
You need a shell on the computer: SSH from the couch. Connect fails? when SSH won't connect. Skip this if you are about to wipe a Marketplace box tonight and reinstall anyway; wipe first, then check health on the fresh install: wipe a Marketplace PC before you install. Don't buy a NAS because one SMART line looked scary before you have a second copy of the house folder.
What this is
Hard drives and many SSDs keep a health log called SMART. The drive counts reallocated sectors, temperature, and other wear signals. A short command asks "are you failing?" You get PASSED, FAILED, or a message that SMART is not available. That answer is not a crystal ball. It is an early warning before photos live only on one dying platter. It is not a reason to skip a backup.
Step 1: install smartmontools
On Ubuntu Server or Debian:
sudo apt update
sudo apt install -y smartmontools
Confirm the tool is there:
smartctl --version
You want a version line, not "command not found."
Step 2: name the disk
List disks without digging into partitions:
lsblk -d -o NAME,SIZE,MODEL,TRAN,TYPE
Pick the main disk, usually sda, nvme0n1, or mmcblk0 on a cheap laptop. Skip loop devices. Skip the USB stick that is only a backup mount. Size and model should match what you expect for the computer. A fuller disk list is see disks and USB sticks on the computer.
Step 3: ask for the overall health
For a SATA or USB-SATA disk named sda:
sudo smartctl -H /dev/sda
For an NVMe SSD named nvme0n1:
sudo smartctl -H /dev/nvme0n1
Look for PASSED near SMART overall-health. FAILED means start a second copy of anything you care about today, then plan a drive swap. "Unable to detect" or "SMART support is unavailable" often means a USB bridge, a cheap SD card, or a virtual disk that does not expose SMART. That is not the same as FAILED.
Step 4: read a little more when PASSED feels thin
Print the short info block:
sudo smartctl -i /dev/sda
Then attributes (SATA style):
sudo smartctl -A /dev/sda
NVMe often wants:
sudo smartctl -A /dev/nvme0n1
You do not need to decode every row. Watch for rising reallocated / pending sectors on spinning disks, and for media or available-spare warnings on NVMe. Temperature in the 30–50 °C range on a quiet shelf is normal. A drive that already FAILED overall health does not need a deep attribute study before you back up.
Step 5: optional short self-test
A short test asks the drive to exercise itself. It can take a few minutes. Start it:
sudo smartctl -t short /dev/sda
Wait the minutes it prints, then:
sudo smartctl -l selftest /dev/sda
Completed without error is good. Aborted or failed means treat the disk as untrusted for the only copy of the house folder. Skip a long test on day one unless you have hours and a second copy already.
If SMART is missing
USB enclosures and some cheap adapters hide SMART. Try:
sudo smartctl -d sat -H /dev/sdX
Swap sdX for the real name. Still nothing? Size and free room still matter (how big, how full), and a second copy still matters more than a perfect health line. Don't buy a monitoring appliance for this.
What this is not
This is not a full disk wipe. It is not a reason to skip a backup of the house folder. It is not RAID. It is not buying a NAS because reallocated sectors ticked up once. It is not replacing free-space checks when copies fail for "no space" reasons.
What next
Make a second copy before the drive gets interesting: a backup of the house folder. Room check: see if the computer still has room. Full disk symptoms: when the computer's disk is full. Still shopping a used box? Wipe first: wipe a Marketplace PC before you install. Size of what you already own: see how big the disk already is.
Optional: jot the model and SMART result in Lab plan — username + passkey, no email. Skip if you just want to keep reading.
Skip
- Skip buying a NAS because one attribute looked weird.
- Skip treating "SMART unavailable" on a USB bridge as FAILED.
- Skip a long self-test before you have any second copy of the house folder.
- Skip wiping the only good backup to "test" the live disk.
- Skip opening the case to swap a drive on day one if PASSED and room is fine.
- Skip RAID as a substitute for a second copy.
Sources
smartmontools. Debian: smartctl. Ubuntu: smartctl. Official pages only. For disk size, see see how big the disk already is. For free room, see see if the computer still has room. For a packed system disk, see when the computer's disk is full. For a second copy, see a backup of the house folder.