Chapter 7: Disks, Partitions & LVM

1. Introduction

Linux stores application data on physical or virtual storage devices. Before creating filesystems or Logical Volumes, it is important to understand how disks are organized and presented to the operating system.

2. Linux Storage Architecture

Physical Disk ──► Partition Table ──► Partitions ──► Filesystem ──► Mount Point ──► Applications
              

Each layer builds upon the previous one, allowing Linux to organize and access persistent storage efficiently.

3. Linux Storage Device Naming

/dev/sda
/dev/sdb
/dev/nvme0n1
/dev/vda

4. Partition Tables

GPT is the preferred partition table format for modern Linux systems because it supports larger disks and more partitions.

5. Common Linux Filesystems

6. Viewing Storage Devices

lsblk
blkid
fdisk -l
parted -l

7. Understanding Mount Points

Disk ──► Partition ──► Filesystem ──► Mounted on ──► /data
              

A mount point is the directory where a filesystem becomes accessible within the Linux directory tree.

8. Linux vs IBM AIX

9. Hands-on Practice

1. List block devices using lsblk.
2. Display filesystem UUIDs using blkid.
3. View partition information using fdisk -l.
4. Identify mounted filesystems.
5. Determine whether your system uses GPT or MBR.

10. Common Mistakes

• Confusing disks with partitions.
• Formatting the wrong device.
• Ignoring UUIDs for persistent mounts.
• Assuming every filesystem behaves the same.

11. Part 1 Summary

12. Creating Partitions with fdisk

fdisk is commonly used to create and manage MBR and GPT partitions on Linux systems. It is interactive: you launch it against a whole disk and type single-letter commands at its prompt.

# Verify the disk name before editing (example uses /dev/sdb — replace with yours)
lsblk
sudo fdisk -l /dev/sdb
sudo fdisk /dev/sdb
m # help: list all commands
p # print the current partition table
n # create a new partition
w # write changes to disk and exit
q # quit without saving

13. Partitioning with parted

parted is recommended for modern GPT disks and large storage devices.

lsblk
sudo parted /dev/sdb print   # inspect first
sudo parted /dev/sdb
print
mklabel gpt                  # DESTRUCTIVE: creates a new GPT label
mkpart data 1MiB 100%
quit

14. Creating Filesystems (mkfs)

Refresh the partition table first

After you write a new partition table, the kernel may still hold its old view of the disk, so the new partition device (for example /dev/sdb1) might not exist yet. Ask the kernel to re-read the table and wait for the device nodes to appear before running mkfs:

sudo partprobe /dev/sdb   # tell the kernel to re-read the partition table
sudo udevadm settle       # wait until the /dev entries are fully created
lsblk /dev/sdb            # confirm the new partition is now listed

Choose one filesystem per partition

Pick a single filesystem for the partition. The commands below are alternatives — you run one of them, not all of them. Running a second mkfs on the same device would overwrite the first filesystem and destroy its data.

sudo mkfs.ext4 /dev/sdb1   # option A: format as ext4
sudo mkfs.xfs  /dev/sdb1   # option B (instead of A): format as XFS
sudo mkfs -t ext4 /dev/sdb1 # equivalent to mkfs.ext4, choosing the type with -t

15. Removing Filesystem Signatures (wipefs)

sudo wipefs /dev/sdb1        # SAFE: only lists the signatures found, changes nothing
sudo wipefs -a /dev/sdb1     # DESTRUCTIVE: erases all filesystem signatures
sudo wipefs --all /dev/sdb1  # same as -a (long form)

Run bare wipefs first: with no options it only lists the filesystem, partition-table, and RAID signatures it detects and makes no changes. Adding -a (or --all) erases those signatures so the OS no longer treats the device as that filesystem. This is not a secure wipe — underlying blocks usually remain until overwritten.

16. Filesystem Repair: Covered in Chapter 9

This chapter focuses on creating filesystems and expanding LVM storage. For unmount requirements, backups, ext4 fsck, and XFS xfs_repair procedures, continue to Chapter 9: Persistent Mounts & Storage Operations.

17. Mounting Filesystems (Quick Look)

sudo mount /dev/sdb1 /data
sudo umount /data

Temporary mounts are useful for testing. For boot-persistent mounts by UUID, fstab options, swap, NFS, and mount troubleshooting, continue to Chapter 9: Persistent Mounts & Storage Operations.

18. Logical Volume Manager (LVM)

Disk ──► Physical Volume (PV) ──► Volume Group (VG) ──► Logical Volume (LV) ──► Filesystem ──► Mount Point
              

LVM provides flexible storage allocations, allowing online filesystem growth. Common setup commands:

lsblk
wipefs /dev/sdb1              # list signatures only — no changes
sudo pvcreate /dev/sdb1       # create a physical volume (destructive to prior FS)
sudo vgcreate vgdata /dev/sdb1
sudo lvcreate -L 20G -n lvdata vgdata

Growing a volume is a two-step job: first enlarge the logical volume, then grow the filesystem that sits on it. The filesystem step differs by type — resize2fs takes the block device, while xfs_growfs takes the mount point:

# Step 1: extend the logical volume by 10 GiB
sudo lvextend -L +10G /dev/vgdata/lvdata

# Step 2a: for an ext4 filesystem, grow it using the device path
sudo resize2fs /dev/vgdata/lvdata

# Step 2b: for an XFS filesystem, grow it using the mount point instead
sudo xfs_growfs /data

You can also let lvextend resize the filesystem in one step with --resizefs:

sudo lvextend -r -L +10G /dev/vgdata/lvdata   # -r is short for --resizefs

The --resizefs option calls the right resize helper for you, but it only supports filesystems lvextend knows how to grow (such as ext4 and XFS) and the filesystem must already exist. When in doubt, run the two steps manually so you can see each result.

19. Common Production Storage Problems

Commands Covered in This Chapter

• lsblk — list block devices
• blkid — show UUIDs and filesystem types
• fdisk / parted — create and manage partitions
• mkfs — create a filesystem
• wipefs — remove filesystem signatures
• mount / umount — mount or unmount a filesystem
• pvcreate / vgcreate / lvcreate — create LVM physical volumes, volume groups, and logical volumes
• lvextend / xfs_growfs / resize2fs — grow logical volumes and filesystems online
• df / du — check disk and directory usage