Learn Rocky Linux - Storage, LVM & Filesystems

#Introduction

In the previous episode 7, you managed services with systemd. But a service without data is just an empty shell — and data must be stored somewhere well-managed. This episode takes you into the world of storage: understanding attached disks, creating partitions and filesystems, mounting them to directories, then leveling up with LVM for expansion flexibility. In episode 3, you chose the LVM layout during installation; now it's time to understand why that choice is so valuable — and how to change it dynamically, without stopping services, when storage needs grow.

#Basic Storage Management

#Mapping Disks

lsblk

blkid

lsblk shows the disk structure as a tree — /dev/sda with partitions sda1, sda2, and so on, including mount points and sizes. blkid shows the UUID and filesystem of each partition.

#Creating Partitions with Parted

To reorganize a disk, you use parted. Modern partitions use the GPT scheme, which supports up to 128 partitions and disks over 2 TB:

parted /dev/sdb --script mklabel gpt
parted /dev/sdb --script mkpart primary ext4 1MiB 100%
partprobe /dev/sdb

parted /dev/sdb print

mklabel gpt creates a GPT partition table, then mkpart creates a new partition from the 1 MiB offset to 100% of the disk. partprobe tells the kernel that the partition table has changed.

#Creating Filesystems and Mounting

mkfs.xfs /dev/sdb1
mkfs.ext4 /dev/sdc1

mkdir -p /mnt/data
mount /dev/sdb1 /mnt/data

XFS is the default RHEL/Rocky filesystem — designed for large scale and high performance. ext4 is a conservative, widely known, and reliable choice.

#Permanent Mounts with fstab

To make a mount survive a reboot, register it in /etc/fstab. Since device names can change, use the UUID:

blkid /dev/sdb1

UUID=abc12345-6789-4def-abcd-1234567890ab /mnt/data xfs defaults,noatime 0 2

The last column 0 2 indicates: not backed up, and the filesystem is checked at boot. Verify the fstab configuration before rebooting:

mount -a
df -h

mount -a tries to mount all fstab entries — if there are no errors, you're safe to reboot.

#Logical Volume Manager (LVM)

#The PV, VG, and LV Concept

LVM is an abstraction layer that separates physical storage from logical storage. Its three core components:

• Physical Volume (PV) — a physical partition or disk handed over to LVM.
• Volume Group (VG) — a collection of PVs pooled into a single storage reservoir.
• Logical Volume (LV) — a logical volume carved out of the VG; this is what gets formatted and mounted.

pvcreate /dev/sdb1
vgcreate vgdata /dev/sdb1
lvcreate -L 50G -n lvdata vgdata

mkfs.xfs /dev/vgdata/lvdata
mount /dev/vgdata/lvdata /mnt/data

LVM's biggest advantage shows when storage runs out. Adding capacity no longer depends on the physical partition size:

pvcreate /dev/sdc1
vgextend vgdata /dev/sdc1

lvextend -L +20G /dev/vgdata/lvdata
xfs_growfs /mnt/data

lvextend grows the LV, then xfs_growfs expands the XFS filesystem to the new size — all without unmounting or downtime.

#LVM Snapshots

A snapshot is a space-efficient point-in-time copy of an LV. Ideal for consistent backups or quick rollback before risky changes:

lvcreate -L 5G -s -n lvdata-snap /dev/vgdata/lvdata

mkdir -p /mnt/snap
mount /dev/vgdata/lvdata-snap /mnt/snap

umount /mnt/snap
lvremove /dev/vgdata/lvdata-snap

A snapshot captures the LV's state at creation time. Backups from a snapshot are more consistent than backing up files that are constantly changing.

#Thin Provisioning

Thin provisioning lets you create LVs that exceed the VG's physical capacity — space is allocated only when data is actually written:

lvcreate -L 100G -T vgdata/thinpool
lvcreate -V 50G -T vgdata/thinpool -n lvthin

#Modern Filesystems

#XFS

As the RHEL/Rocky default, XFS excels with large files, high throughput, and scalability. It doesn't support shrinking, but its reliability and performance make it the top choice for production workloads like databases and file servers.

#Btrfs

Btrfs brings modern features like subvolumes and built-in snapshots:

mkfs.btrfs /dev/sdd1
mount /dev/sdd1 /mnt/data
btrfs subvolume create /mnt/data/@snap

Btrfs snapshots are integrated with the filesystem itself, unlike LVM snapshots which live at a lower layer.

#Stratis

Stratis is a storage manager layered on top of LVM and XFS that presents a simple storage concept:

dnf5 install stratisd stratis-cli
systemctl enable --now stratisd
stratis pool create pool1 /dev/sde1
stratis fs create pool1 data
stratis fs list

Stratis delivers enterprise-level features — pooling, snapshots, and automatic expansion — with far simpler commands than manually managing LVM. It's available on Rocky 9 and 10 for those who want a higher-level abstraction.

#Choosing an Approach

A quick guide to choosing your storage toolkit:

• Simple and proven: direct partitions with XFS or ext4.
• Flexibility and dynamic expansion: LVM with PV/VG/LV.
• Built-in snapshots and subvolumes: Btrfs.
• Ease of layered management: Stratis.

#Closing

In this episode 8, you mastered Rocky Linux storage management: disk mapping with lsblk and blkid, GPT partition creation with parted, XFS and ext4 filesystem creation, permanent mounts with UUID-based fstab, LVM with PV/VG/LV including downtime-free expansion and snapshots, thin provisioning, and the modern Btrfs and Stratis filesystems.

Key takeaways:

• Always use lsblk and blkid to understand the storage situation before acting.
• Use UUIDs in /etc/fstab and test with mount -a before rebooting.
• LVM allows storage expansion without stopping services via lvextend and xfs_growfs.
• LVM snapshots provide consistent point-in-time backups for quick rollback.
• XFS is a reliable default; Btrfs and Stratis offer modern features for specific needs.

In the next episode 9, we will discuss networking and hostname — network configuration with NetworkManager and nmcli, static IP and DHCP connection profiles, bonding, VLANs, and bridges, plus routes and hostname. Data is stored safely; now it's time to make your system communicate with the outside world!