Day 11 - Disks and Filesystems

This lesson builds a mental model of Linux storage and teaches the daily tools for working with disks and filesystems. It uses a virtual machine example so the workflow is safe to practice.

Prerequisites

• Day 1 through Day 10 completed
• A VM where an extra virtual disk can be added, or permission to create a loopback device

Inspect devices and layout

1
2
3
4
5
6
7

lsblk -o NAME,TYPE,SIZE,FSTYPE,MOUNTPOINT,LABEL,UUID

# tree view of mounts by source and target
findmnt -A | head -30

# kernel messages for storage
sudo dmesg | grep -iE 'sd[a-z]|nvme|virtio' | tail -20

Typical device names:

• SATA or SCSI: /dev/sda, /dev/sdb and partitions /dev/sda1, /dev/sda2
• NVMe: /dev/nvme0n1 and partitions /dev/nvme0n1p1
• Virtual disks under KVM or QEMU: vda, vdb

Add a new virtual disk in a VM (safe path)

1. Shut down the VM and add a second disk in the hypervisor UI, for example 5 GB.
2. Boot and confirm it appears as an unpartitioned device.

1

lsblk -o NAME,TYPE,SIZE,FSTYPE | grep -E 'disk|part'

If adding a real disk is not possible, create a loopback device from a file for practice.

1
2
3
4

# create a 2 GiB sparse file as a demo disk
sudo fallocate -l 2G /var/tmp/disk.img
sudo losetup --find --show /var/tmp/disk.img   # prints /dev/loopX
lsblk -o NAME,TYPE,SIZE /dev/loop*

Partition the disk

Use parted for a quick guided setup. GPT is recommended for modern systems.

1
2
3
4

DISK=/dev/sdb   # replace with your device or the loop device
sudo parted -s $DISK mklabel gpt
sudo parted -s $DISK mkpart primary ext4 1MiB 100%
sudo parted -s $DISK print

The new partition appears as $DISK1 such as /dev/sdb1 or /dev/loop0p1.

Alternative with fdisk (interactive):

1
2

sudo fdisk /dev/sdb
# n (new) p (primary) 1, accept defaults, w (write)

Make a filesystem and mount it

Create an ext4 filesystem and a mount point.

1
2
3
4
5
6
7
8

PART=/dev/sdb1   # adjust to match your partition
sudo mkfs.ext4 -L data1 $PART
sudo mkdir -p /mnt/data1
sudo mount $PART /mnt/data1

# confirm
lsblk -o NAME,SIZE,FSTYPE,MOUNTPOINT $PART
findmnt /mnt/data1

Test write access.

1
2

sudo sh -c 'echo test > /mnt/data1/hello.txt'
sudo ls -l /mnt/data1

Unmount when done or before making it persistent.

1

sudo umount /mnt/data1

Persistent mounts with /etc/fstab

Mounts added at runtime vanish after reboot. Use /etc/fstab for persistence, preferably by UUID or label.

Find the UUID and label.

1

blkid $PART

Add a line to /etc/fstab.

1
2
3
4
5
6
7
8
9
10
11

# backup first
sudo cp /etc/fstab /etc/fstab.bak

# example using UUID and ext4 defaults
UUID=$(blkid -s UUID -o value $PART)
LINE="UUID=$UUID /mnt/data1 ext4 defaults,nofail 0 2"
echo "$LINE" | sudo tee -a /etc/fstab

# test without reboot
sudo mount -a
findmnt /mnt/data1

Notes:

• nofail allows boot to continue if the device is not present.
• The final 2 enables fsck order. Root is usually 1, others 2.

Space and usage checks

1
2
3
4
5
6
7
8

# filesystem space and type
df -hT

# disk usage by directory, current level
du -sh * | sort -h

# restrict to one filesystem and go two levels deep
du -h -x --max-depth=2 /var | sort -h | tail -20

1

sudo find / -xdev -type f -size +1G -printf '%p %s bytes\n' 2>/dev/null | head

Labels and tuning

Labels and tunables help identify and optimize filesystems.

1
2
3
4
5

# set or change an ext4 label
sudo e2label $PART projects

# show and tweak ext4 parameters
tune2fs -l $PART | head -20

Mount options influence performance and safety.

• noatime reduces write amplification on frequently read files
• nodev, nosuid, noexec reduce risk on untrusted media

Example temporary mount with options:

1

sudo mount -o noatime,nodev /dev/sdb1 /mnt/data1

Persist options by adding them in the fstab options column.

Create and manage a swapfile (optional)

A swapfile is flexible and avoids repartitioning.

1
2
3
4
5
6
7
8
9
10
11
12

# create a 2 GiB swapfile
sudo fallocate -l 2G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile

# persist across reboots
echo '/swapfile none swap sw 0 0' | sudo tee -a /etc/fstab

# verify
swapon --show
free -h

Check and repair filesystems

For ext filesystems, fsck checks integrity when a filesystem is unmounted.

1
2

sudo umount /mnt/data1
sudo fsck -f $PART

Do not run fsck on a mounted, writable filesystem.

Practical lab

1. Create a new partition and ext4 filesystem on a secondary disk or loop device, then mount it at /mnt/data1.

1
2
3
4
5
6
7
8

# substitute your device
DISK=/dev/sdb
PART=/dev/sdb1
sudo parted -s $DISK mklabel gpt
sudo parted -s $DISK mkpart primary ext4 1MiB 100%
sudo mkfs.ext4 -L data1 $PART
sudo mkdir -p /mnt/data1
sudo mount $PART /mnt/data1

2. Make the mount persistent by UUID with safe options.

1
2
3
4

UUID=$(blkid -s UUID -o value $PART)
echo "UUID=$UUID /mnt/data1 ext4 defaults,nofail 0 2" | sudo tee -a /etc/fstab
sudo mount -a
findmnt /mnt/data1

3. Generate space and measure it.

1
2
3

sudo dd if=/dev/zero of=/mnt/data1/bigfile bs=100M count=5 status=progress
sudo du -sh /mnt/data1
sudo df -hT | grep data1

4. Optional: create a 1 GiB swapfile and verify it is active.

1
2
3
4
5

sudo fallocate -l 1G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile
swapon --show

5. Clean up the practice environment if using a loop device.

1
2
3
4

sudo swapoff /swapfile && sudo rm -f /swapfile || true
sudo umount /mnt/data1 || true
sudo losetup -D || true
sudo rm -f /var/tmp/disk.img || true

Troubleshooting

• device is busy when unmounting. Close terminals, stop services using the mount, or use lsof +f -- /mnt/data1 to find holders.
• unknown filesystem type after creating a filesystem. Confirm the correct mkfs tool was used and that the kernel supports the type.
• Mount fails on boot. Use nofail in /etc/fstab, verify UUIDs, and test with mount -a. Check logs with journalctl -b -u systemd-fsck and journalctl -b.
• Permission issues on shared data directories. Use correct ownership and group settings, for example sudo chown -R root:developers /mnt/data1/projects && sudo chmod -R 2775 /mnt/data1/projects.
• Swapfile cannot be enabled. Ensure file mode is 600 and that a sparse file is not on a filesystem with nocow options that block swap on some filesystems.

Up next: Day 12 - Finding Things and Text Processing I

Day 12 begins text processing with grep, cut, sort, uniq, wc, and tee. It focuses on searching and shaping logs and command output into quick reports.