02 - Storage and Filesystems¶

Block devices¶

lsblk                  # tree view of disks, partitions, mounts
fdisk -l               # detailed partition table
blkid                  # UUID + filesystem type

Common naming: /dev/sda, /dev/nvme0n1, /dev/vdb (virtio).

Partitions¶

MBR (legacy) vs GPT¶

MBR - max 4 primary partitions, 2 TB limit. Use for legacy BIOS systems.
GPT - many partitions, no size limit. Default on modern UEFI systems.

Create a partition with `fdisk`¶

fdisk /dev/sdb
# n  - new partition
# p  - primary (or e for extended)
# 1  - partition number
# (Enter)  - default first sector
# +5G  - size
# t  - change type
# 8e - Linux LVM (or 82 for swap, 83 for filesystem)
# w  - write

After creating partitions:

partprobe /dev/sdb     # rescan partition table

Or with `parted` (scriptable)¶

parted /dev/sdb mklabel gpt
parted /dev/sdb mkpart primary xfs 0% 5GB

Filesystems¶

Filesystem	Best for
xfs	Default RHEL 9 root and data; can grow but cannot shrink
ext4	Mature, can shrink; good for older workloads
vfat	EFI partition / cross-platform USB
swap	Swap space (not a real filesystem)
iso9660	DVD / ISO images

Create a filesystem¶

mkfs.xfs /dev/sdb1
mkfs.ext4 -L mylabel /dev/sdb2
mkswap /dev/sdb3

Mounting¶

Manual mount¶

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

Persistent mount (`/etc/fstab`)¶

Format:

<device>    <mount_point>    <fstype>    <options>    <dump>    <fsck>

Example:

UUID=abc-123    /mnt/data    xfs    defaults    0    0
/dev/vg_data/lv_data    /var/lib/postgres    ext4    defaults,noatime    0    2
LABEL=swap1    none    swap    defaults    0    0

Always use UUID (or LABEL) instead of device name (/dev/sdX1) - device names can change between reboots.

Get UUID:

blkid /dev/sdb1
# /dev/sdb1: UUID="..." BLOCK_SIZE="512" TYPE="xfs" PARTUUID="..."

After editing fstab, always test before reboot:

systemctl daemon-reload   # if any systemd-managed mounts
mount -a                  # mounts everything in fstab not already mounted

If mount -a errors, fix fstab before rebooting - otherwise the system may fail to boot.

LVM (Logical Volume Manager)¶

LVM gives you flexible storage that can grow / shrink without re-partitioning.

Layers¶

Physical disk(s)  →  Physical Volume (PV)  →  Volume Group (VG)  →  Logical Volume (LV)  →  Filesystem

Create LVM stack¶

# Mark partitions as LVM (type 8e via fdisk)
pvcreate /dev/sdb1 /dev/sdc1
pvs                                       # list PVs

vgcreate vg_data /dev/sdb1 /dev/sdc1
vgs                                       # list VGs

lvcreate -L 5G -n lv_data vg_data         # 5 GB logical volume
lvcreate -l 100%FREE -n lv_logs vg_data   # use all remaining space
lvs                                       # list LVs

mkfs.xfs /dev/vg_data/lv_data
mkdir -p /mnt/data
mount /dev/vg_data/lv_data /mnt/data

Then add to fstab using UUID.

Extend an LV (online, no downtime)¶

# Add space to VG first if needed
vgextend vg_data /dev/sdd1

# Extend LV
lvextend -L +2G /dev/vg_data/lv_data

# Or grow to use all free VG space
lvextend -l +100%FREE /dev/vg_data/lv_data

# Resize the filesystem
xfs_growfs /mnt/data         # XFS (online)
resize2fs /dev/vg_data/lv_data  # ext4 (online)

Reduce an LV (ext4 only - XFS cannot shrink)¶

umount /mnt/data
e2fsck -f /dev/vg_data/lv_data
resize2fs /dev/vg_data/lv_data 4G
lvreduce -L 4G /dev/vg_data/lv_data
mount /mnt/data

Stratis (RHEL 9)¶

Stratis is a higher-level storage manager built on LVM and XFS. The exam touches it lightly.

dnf install -y stratis-cli stratisd
systemctl enable --now stratisd

stratis pool create mypool /dev/sdb /dev/sdc
stratis filesystem create mypool myfs

# Mount path:
mount /dev/stratis/mypool/myfs /mnt/strat

Persistent mount uses /dev/stratis/<pool>/<fs> and the xfs filesystem with x-systemd.requires=stratisd.service option.

Swap¶

Create a swap partition¶

mkswap /dev/sdb3
swapon /dev/sdb3
swapon --show

Add to fstab:

UUID=...  none  swap  defaults  0  0

Create a swap file¶

dd if=/dev/zero of=/swapfile bs=1M count=2048
chmod 600 /swapfile
mkswap /swapfile
swapon /swapfile
echo '/swapfile  none  swap  defaults  0  0' >> /etc/fstab

NFS (network filesystem)¶

Server side (not always asked)¶

dnf install -y nfs-utils
echo '/srv/share  192.168.1.0/24(rw,sync,no_root_squash)' >> /etc/exports
systemctl enable --now nfs-server
exportfs -rav
firewall-cmd --permanent --add-service=nfs --add-service=mountd --add-service=rpc-bind
firewall-cmd --reload

Client side (more often tested)¶

dnf install -y nfs-utils
mkdir -p /mnt/nfs
mount -t nfs server:/srv/share /mnt/nfs

Persistent in /etc/fstab:

server:/srv/share   /mnt/nfs   nfs   defaults,_netdev   0   0

_netdev ensures the mount waits for network availability at boot.

Automounting NFS¶

autofs mounts on access and unmounts on idle.

dnf install -y autofs
systemctl enable --now autofs

/etc/auth.master.d/share.autofs:

/-  /etc/auto.share

/etc/auto.share:

/mnt/share  -fstype=nfs,rw  server:/srv/share

ACLs (Access Control Lists)¶

Standard Unix permissions limit you to user/group/other. ACLs allow per-user / per-group rules.

# Set ACL
setfacl -m u:alice:rw- /path/file
setfacl -m g:devs:rwx /path/dir

# View ACLs
getfacl /path/file

# Default ACL on dir (applied to new children)
setfacl -d -m u:alice:rw- /path/dir

# Remove
setfacl -x u:alice /path/file

# Remove all ACLs
setfacl -b /path/file

Files with ACLs show a + in ls -l:

-rw-rw-r--+ 1 root root  ...  file

Common tasks and patterns¶

Add a 5 GB partition, format ext4, mount at /var/log/app, persistent¶

# 1. Create partition
fdisk /dev/sdb     # n, default size +5G, type 83, w
partprobe

# 2. Create filesystem
mkfs.ext4 -L applogs /dev/sdb1

# 3. Get UUID
blkid /dev/sdb1

# 4. Add to /etc/fstab
echo 'UUID=<uuid>  /var/log/app  ext4  defaults  0  2' >> /etc/fstab

# 5. Test
mkdir -p /var/log/app
mount -a
df -h | grep applogs

# 6. Reboot to verify
reboot

Add a 5 GB LV in vg_data, mount at /opt/myapp¶

lvcreate -L 5G -n lv_myapp vg_data
mkfs.xfs /dev/vg_data/lv_myapp
mkdir -p /opt/myapp
echo '/dev/vg_data/lv_myapp  /opt/myapp  xfs  defaults  0  2' >> /etc/fstab
mount -a
df -h | grep lv_myapp

Extend an existing logical volume to use all remaining space in VG¶

lvextend -l +100%FREE /dev/vg_data/lv_data
xfs_growfs /mnt/data
df -h

Configure a swap file of 1 GB¶

dd if=/dev/zero of=/swapfile bs=1M count=1024
chmod 600 /swapfile
mkswap /swapfile
swapon /swapfile
echo '/swapfile  none  swap  defaults  0  0' >> /etc/fstab
swapon --show

Verification checklist¶

lsblk shows the new device with the right mount point
df -h shows the filesystem
mount shows mount options as expected
cat /proc/swaps for swap
After reboot, all of the above still hold