How to Check a Disk for Bad Blocks and Errors in Linux

Updated: 30 September 2026 · Applies to: smartmontools 7.2 to 7.4, e2fsprogs badblocks and xfsprogs on AlmaLinux and Rocky Linux 9 and 10, RHEL 9 and 10, Ubuntu 24.04 and 26.04, Debian 13

Disk errors show up as I/O errors in the logs, slow file access, or a file system that turns read-only. Modern disks (HDD and SSD) replace bad sectors themselves and report their health through S.M.A.R.T., so that is the first thing to check. Scanning the surface with badblocks is a second step, mainly for hard disks. We ran the commands on AlmaLinux 9.

1. Look for errors in the logs

journalctl -k -p err | tail -30
dmesg -T | grep -i -E "i/o error|medium error|ata[0-9]+.*error|nvme.*error"

2. Check the disk's own health (S.M.A.R.T.)

dnf install smartmontools     # or: apt install smartmontools
smartctl -H /dev/sda          # overall result: PASSED or FAILED
smartctl -a /dev/sda          # all attributes and the error log
smartctl -t short /dev/sda    # start a short self-test (about 2 minutes)
smartctl -l selftest /dev/sda # result of the self-tests

Warning signs on hard disks are rising values for Reallocated_Sector_Ct, Current_Pending_Sector and Offline_Uncorrectable; on NVMe SSDs, Media and Data Integrity Errors and a high Percentage Used. For details, see how to install and use smartctl and how to monitor disk health with smartd.

On virtual servers, the virtual disks have no S.M.A.R.T. data; disk health is your provider's responsibility there.

3. Scan for bad blocks (hard disks)

A read-only scan is safe on a disk with data, but it takes hours on large disks and should run when the disk is not busy:

badblocks -sv /dev/sdb

-s shows progress and -v reports what it finds; the end result looks like Pass completed, 0 bad blocks found. Never use the write test (-w) on a disk with data: it overwrites everything.

4. Check the file system

The file system must be unmounted (for the root file system, from a rescue system):

xfs_repair -n /dev/sdb1       # XFS: check only, no changes
e2fsck -fn /dev/sdb1          # ext4: check only, no changes

For ext4, e2fsck -c runs a read-only bad block scan and records bad blocks in the file system so they are not used. XFS has no such list; it relies on the disk to remap bad sectors.

What to do when a disk is failing

• Back up the data immediately. A disk with growing reallocated or pending sectors usually gets worse.
• Replace the disk. In a RAID array, replace the failed member and let the array rebuild.
• On a dedicated server, contact your provider with the smartctl -a output so the disk can be swapped.

Official documentation: smartmontools and badblocks manual.