Updated: 28 September 2026 · Applies to: GNU parted 3.5 and 3.6 on AlmaLinux and Rocky Linux 9 and 10, RHEL 9 and 10, Ubuntu 24.04 and 26.04, Debian 13

This guide takes a new, empty disk (for example a second disk added to a dedicated server or VPS) from nothing to a mounted folder that survives reboots: partition it with parted, create a file system, and add it to /etc/fstab. For more parted commands, such as resizing and recovering partitions, see 8 Linux parted commands.

Before you start

  • Every step below erases the disk you name. Make sure it is the new, empty disk.
  • Install parted if it is missing: dnf install parted or apt install parted.

Steps

  1. Find the new disk. It has no partitions and no mount point:
    lsblk -o NAME,SIZE,TYPE,FSTYPE,MOUNTPOINTS
    In this example it is /dev/sdb. On a VPS it may be /dev/vdb, on NVMe servers /dev/nvme1n1.
  2. Create a GPT partition table and one partition over the whole disk:
    parted -s /dev/sdb mklabel gpt mkpart data 1MiB 100%
    GPT works for disks of any size (MBR stops at 2 TB). Starting at 1 MiB keeps the partition aligned.
  3. Check the result: parted /dev/sdb print. The partition is /dev/sdb1 (on NVMe: /dev/nvme1n1p1).
  4. Create a file system. XFS is the default on AlmaLinux, Rocky Linux and RHEL, ext4 on Ubuntu and Debian; both are good choices:
    mkfs.xfs /dev/sdb1       # or: mkfs.ext4 /dev/sdb1
  5. Create the folder and mount it:
    mkdir -p /data
    mount /dev/sdb1 /data
    df -h /data
  6. Get the file system's UUID, which stays the same even if disk names change:
    blkid /dev/sdb1
  7. Add a line to /etc/fstab with that UUID:
    UUID=0a1b2c3d-1111-2222-3333-444455556666  /data  xfs  defaults,nofail  0 0
    Use ext4 instead of xfs if you chose ext4. nofail lets the server boot even if this disk is missing.
  8. Check the fstab line before you reboot:
    umount /data
    systemctl daemon-reload
    mount -a
    findmnt --verify
    mount -a must mount /data again without errors. A broken fstab line can stop a server from booting, so never skip this check.

Common problems

  • "Partition(s) ... have been written, but we have been unable to inform the kernel": run partprobe /dev/sdb, or reboot, before creating the file system.
  • Wrong disk chosen: stop at once. Data can sometimes be recovered with parted rescue or TestDisk if nothing more was written.
  • Want to use it with LVM instead: skip the file system step and run pvcreate /dev/sdb1, then add it to a volume group.

Official documentation: GNU parted manual and fstab manual page.

Ucartz services for this topic

Was this answer helpful? 0 Users Found This Useful (0 Votes)