RED HAT ENTERPRISE LINUX

Non-Destructive Storage
Expansion

Add new partitions and logical volumes, and swap to a system non-destructively

CIS126RH | RHEL System Administration 1
Mesa Community College

Production servers must grow their storage capacity without downtime. Adding a partition to a disk that already has data, extending an LVM volume group and logical volume while filesystems are mounted, and increasing swap space on a live system are all operations that must leave existing data completely intact. This module integrates the partition, LVM, swap, and fstab skills from previous modules into a complete non-destructive expansion workflow tested on the RHCSA exam.

Learning Objectives

1. Add a new partition to a disk with existing partitions — Use parted to create a new partition in free space on a disk that already has data partitions, without touching existing partitions
2. Extend an LVM volume group and logical volume online — Add a new physical volume to an existing VG and extend a logical volume and its filesystem while it remains mounted
3. Add swap space non-destructively — Create a new swap partition or LVM swap volume, activate it, and make it persistent in /etc/fstab
4. Verify all changes persist after reboot — Confirm that new mounts, extended filesystems, and swap are all correctly configured in fstab and survive a system restart

What "Non-Destructive" Means

Every storage expansion operation on a running system must preserve all existing data. "Non-destructive" imposes specific constraints on each operation.

Adding a Partition to a Disk with Existing Data

The disk must have free (unpartitioned) space. Verify the layout first, then create a new partition in the available space.

# Step 1: Inspect the current partition layout
$ sudo parted /dev/sdb print
Model: ATA VBOX HARDDISK (scsi)
Disk /dev/sdb: 20.0GB
Partition Table: gpt

Number  Start    End      Size    File system  Name     Flags
 1      1049kB   5369MB   5368MB  xfs          primary
                                    ← 14.6 GB free after partition 1

# Step 2: Create a new partition in the free space
$ sudo parted /dev/sdb
(parted) mkpart primary xfs 5369MB 10737MB   # 5 GB partition
(parted) print
Number  Start    End      Size    File system  Name
 1      1049kB   5369MB   5368MB  xfs          primary   ← untouched
 2      5369MB   10737MB  5368MB  xfs          primary   ← new
(parted) quit

# Step 3: Update the kernel's view
$ sudo partprobe /dev/sdb
$ lsblk /dev/sdb

Format and Mount the New Partition

After creating the partition and updating the kernel, format and mount it, then add the persistent fstab entry using UUID.

# Format the new partition (existing partition is NOT touched)
$ sudo mkfs.xfs /dev/sdb2
meta-data=/dev/sdb2  isize=512  agcount=4...

# Get the UUID of the new filesystem
$ sudo blkid /dev/sdb2
/dev/sdb2: UUID="mnp012-qr34-st56-uv78-wx901234yzab" TYPE="xfs"

# Create the mount point
$ sudo mkdir -p /newdata

# Add the persistent fstab entry
$ echo "UUID=mnp012-qr34-st56-uv78-wx901234yzab  /newdata  xfs  defaults  0 2" | \
  sudo tee -a /etc/fstab

# Mount and verify
$ sudo mount -a
$ df -h /newdata

# Verify the existing partition is still intact
$ ls /existingdata   # existing data is untouched

Extending an LVM Volume Group Online

Adding capacity to an existing LVM setup is a live, non-disruptive operation. The volume group, logical volume, and filesystem all grow while mounted.

# Step 1: Confirm current state — VG is nearly full
$ sudo vgs
  VG     #PV #LV Attr   VSize   VFree
  datavg   1   1 wz--n- <10.00g 500.00m

# Step 2: Prepare new storage (partition + PV)
$ sudo parted --script /dev/sdc mklabel gpt mkpart primary 1MiB 100%
$ sudo partprobe /dev/sdc
$ sudo pvcreate /dev/sdc1

# Step 3: Add the PV to the existing VG
$ sudo vgextend datavg /dev/sdc1
  Volume group "datavg" successfully extended

# Step 4: Extend the LV and grow the filesystem at once
$ sudo lvextend -L +8G -r /dev/datavg/data
  Size of logical volume datavg/data changed to 10.00 GiB
  Filesystem on /dev/datavg/data resized successfully

# Step 5: Verify — LV and filesystem both grew
$ sudo lvs datavg/data
$ df -h /data

Adding a New LV to an Existing VG

If the VG has free space, a new logical volume can be created immediately without adding any new physical storage.

# Check available free space in the VG
$ sudo vgs
  VG     #PV #LV Attr   VSize   VFree
  datavg   2   1 wz--n- <20.00g <10.00g  ← 10 GB free

# Create a new LV from the VG's free space
$ sudo lvcreate -L 5G -n logs datavg
  Logical volume "logs" created.

# Format, mount, add to fstab
$ sudo mkfs.xfs /dev/datavg/logs
$ sudo mkdir -p /logs
$ sudo blkid /dev/datavg/logs
# Add UUID entry to /etc/fstab
$ sudo mount -a
$ df -h /logs

# Confirm existing LV is still intact and mounted
$ df -h /data
$ sudo lvs

Adding Swap: Partition Method

When free disk space is available, a new swap partition is the simplest way to add swap capacity without affecting existing partitions.

# Step 1: Check current swap
$ free -h
               total        used        free      shared
Swap:          2.0Gi       512Mi       1.5Gi

$ swapon --show
NAME      TYPE      SIZE   USED PRIO
/dev/sda2 partition   2G  512M   -2

# Step 2: Create a new partition for swap
$ sudo parted --script /dev/sdb mkpart primary linux-swap 10737MB 12785MB
$ sudo partprobe /dev/sdb

# Step 3: Format as swap
$ sudo mkswap /dev/sdb3
Setting up swapspace version 1, size = 2 GiB

# Step 4: Activate immediately
$ sudo swapon /dev/sdb3

# Step 5: Add persistent fstab entry using UUID
$ sudo blkid /dev/sdb3
/dev/sdb3: UUID="xyz789-ab01-cd23-ef45-gh6789ijklmn" TYPE="swap"
$ echo "UUID=xyz789-ab01-cd23-ef45-gh6789ijklmn  none  swap  defaults  0 0" | \
  sudo tee -a /etc/fstab

# Step 6: Verify
$ swapon --show

Adding Swap: LVM Method

Creating swap as a logical volume uses existing VG free space and is the preferred method when the system already uses LVM.

# Step 1: Confirm VG has free space
$ sudo vgs
  VG     VSize   VFree
  datavg <20.00g  <5.00g

# Step 2: Create a swap LV from VG free space
$ sudo lvcreate -L 2G -n swap2 datavg
  Logical volume "swap2" created.

# Step 3: Format as swap
$ sudo mkswap /dev/datavg/swap2
Setting up swapspace version 1, size = 2 GiB

# Step 4: Activate immediately
$ sudo swapon /dev/datavg/swap2

# Step 5: Add persistent fstab entry (LVM path is stable)
$ echo "/dev/datavg/swap2  none  swap  defaults  0 0" | \
  sudo tee -a /etc/fstab

# Step 6: Verify — existing swap and new swap both active
$ swapon --show
NAME                TYPE      SIZE USED PRIO
/dev/sda2           partition   2G 512M   -2
/dev/mapper/datavg-swap2 partition 2G   0B   -3

Checking and Managing Swap

Several commands provide visibility into current swap usage and configuration.

# Show all swap devices with size, usage, and priority
$ swapon --show
NAME              TYPE      SIZE   USED PRIO
/dev/sda2         partition   2G  256M   -2
/dev/datavg/swap2 partition   2G    0B   -3

# Show swap summary alongside RAM
$ free -h
               total        used        free
Mem:           7.6Gi       2.1Gi       5.5Gi
Swap:          4.0Gi       256Mi       3.7Gi  ← 4 GiB total (2+2)

# Deactivate a specific swap device (e.g. before removal)
$ sudo swapoff /dev/datavg/swap2

# Activate all swap listed in /etc/fstab
$ sudo swapon -a

# Set higher priority for a swap device (used first)
$ sudo swapon -p 10 /dev/datavg/swap2
# In fstab, use: /dev/datavg/swap2  none  swap  defaults,pri=10  0 0

The Non-Destructive Expansion Checklist

Before and after every storage expansion, follow this checklist to confirm that existing data and mounts are intact.

1. Before any changes — record the baseline
   lsblk, df -h, swapon --show, sudo pvs, sudo vgs, sudo lvs
2. Identify available free space
   Parted: gap after last partition. LVM: VFree in vgs output. Disk: unpartitioned space shown in lsblk.
3. Make changes using only the identified free space
   New partition starts where the last one ended. New PV/LV uses VG free space, not existing allocations.
4. After changes — verify both old and new
   Existing mounts still accessible: ls /existingmount
   Existing swap still active: swapon --show
   New mount working: df -h /newmount
5. Make persistent — add fstab entries
   Test with sudo mount -a and findmnt --verify

Complete Scenario: Add Storage to a Live Server

A server with /dev/sdb (10 GB, has one 5 GB XFS partition at /data) needs a second 3 GB data partition and 1 GB more swap. All without touching the existing /data filesystem.

Part A: New data partition

$ sudo parted /dev/sdb print              # confirm free space starts at 5369MB
$ sudo parted --script /dev/sdb mkpart primary xfs 5369MB 8590MB
$ sudo partprobe /dev/sdb
$ sudo mkfs.xfs /dev/sdb2
# blkid → add UUID to fstab → mount -a → df -h /data2

Part B: New swap partition

$ sudo parted --script /dev/sdb mkpart primary linux-swap 8590MB 9663MB
$ sudo partprobe /dev/sdb
$ sudo mkswap /dev/sdb3
$ sudo swapon /dev/sdb3
# blkid → add UUID swap entry to fstab

Verification

$ ls /data         # existing partition untouched
$ df -h /data2    # new data partition working
$ swapon --show   # both old and new swap active

Complete Scenario: Expand LVM Storage Online

A server's /dev/datavg/data LV mounted at /data is 80% full. A new 10 GB disk has been added. Expand without downtime.

# 1. Baseline
$ df -h /data      # record current size and usage
$ ls /data         # confirm existing data is present

# 2. Prepare new disk as PV
$ sudo parted --script /dev/sdc mklabel gpt mkpart primary 1MiB 100%
$ sudo partprobe /dev/sdc
$ sudo pvcreate /dev/sdc1

# 3. Add PV to existing VG
$ sudo vgextend datavg /dev/sdc1

# 4. Extend LV and filesystem (all online)
$ sudo lvextend -l +100%FREE -r /dev/datavg/data

# 5. Verify — existing data intact, filesystem is larger
$ ls /data         # original files still present
$ df -h /data      # size has increased
$ sudo lvs          # LV shows new size
$ sudo vgs          # VFree should be ~0 (used all free space)

Making All Changes Persistent

New mounts and swap must be in /etc/fstab to survive reboots. Extended LV filesystems survive automatically — the LV size is stored in LVM metadata.

# View the current fstab to avoid duplicate entries
$ cat /etc/fstab

# Add new partition mount (UUID from blkid)
UUID=mnp012-qr34-st56...  /data2   xfs   defaults   0 2

# Add new swap partition (UUID from blkid)
UUID=xyz789-ab01-cd23...  none     swap  defaults   0 0

# Add new LVM swap (device path is stable for LVM)
/dev/datavg/swap2          none     swap  defaults   0 0

# Test all new entries without rebooting
$ sudo mount -a
$ sudo swapon -a

# Verify all mounts and swap are active
$ findmnt --verify
$ df -h
$ swapon --show

# IMPORTANT: Do NOT need to add entry for extended LV
# The existing fstab entry already references /dev/datavg/data
# LVM metadata stores the new size — it persists automatically

Quick Reference: Non-Destructive Expansion

Common Mistakes

Knowledge Check

Answer these before moving to the next slide.

1. A disk has one partition ending at 5369 MB on a 20 GB disk. Write the parted command to add a 3 GB partition starting immediately after the existing one.
2. What does "non-destructive" mean in the context of adding a partition to a disk that already has data?
3. An LV /dev/datavg/data mounted at /data needs 8 GB more space. The VG has 10 GB free. Write the single command to extend the LV and grow the XFS filesystem simultaneously.
4. You have added a new swap partition at /dev/sdb3 and run mkswap. What two things must you do to make the swap (a) active now and (b) persistent across reboots?
5. After extending an LV, do you need to add a new fstab entry for the extended filesystem? Explain why or why not.
6. List the three commands to verify, after all changes, that existing data is still intact and the new storage is working.

Knowledge Check — Answers

1. sudo parted --script /dev/sdb mkpart primary xfs 5369MB 8438MB
   Start: 5369 MB (immediately after the existing partition's end). End: 5369 + 3072 ≈ 8438 MB (approximately 3 GB).
2. Non-destructive means the new partition is created only in free (unpartitioned) space on the disk. The existing partition's start sector, end sector, and all data within it remain completely unchanged. The operation is additive — it adds a new partition without modifying, resizing, or deleting any existing partition.
3. sudo lvextend -L +8G -r /dev/datavg/data
   The -r flag automatically runs xfs_growfs on the mounted XFS filesystem after extending the LV, growing both in one step.
4. (a) Active now: sudo swapon /dev/sdb3
   (b) Persistent: add UUID=... none swap defaults 0 0 (using the UUID from blkid /dev/sdb3) to /etc/fstab.
5. No, a new fstab entry is not needed. The existing fstab entry still references the same LV (e.g. /dev/datavg/data or its UUID). LVM metadata stores the LV's new size persistently — on the next boot, the LV activates at its new size and the filesystem mounts at that larger size using the existing fstab entry.
6. Any three of: ls /existingmount (existing data intact), df -h (all filesystems and sizes visible), df -h /newmount (new mount accessible), swapon --show (all swap active including new), findmnt --verify (fstab syntax valid), sudo lvs (LV sizes).

Key Takeaways

1. Always inspect the current layout before making any changes. Use lsblk, parted print, pvs, vgs, and swapon --show to record the baseline. New partitions start where old ones end. New LVs use only VG free space.
2. Non-destructive means additive — free space only, no modification of existing. New partitions use unpartitioned gaps. New LVs use unallocated VG extents. New swap is additional, not a replacement. Verify existing data after every change.
3. Use lvextend -r to extend an LV and its filesystem in one step. The -r flag detects the filesystem type and runs the correct grow tool. The existing fstab entry survives — do not add a new one for an extended LV.
4. New mounts and swap need fstab entries; extended LVs do not. New partition mounts: UUID= in fstab. New swap: UUID= or device path with type swap. Extended LV: existing fstab entry is sufficient — LVM metadata stores the new size. Test everything with mount -a and findmnt --verify.

Graded Lab

• Run lsblk, df -h, and swapon --show to record the baseline. Identify any disks with free (unpartitioned) space and any VGs with free extents.
• On a disk with free space after its last partition, add a new 2 GiB partition. Run partprobe, format with XFS, add a UUID-based fstab entry, and mount. Verify the previous partition's data is untouched.
• If a VG with free extents exists, create a new 1 GiB logical volume named labextra. Format with XFS, add a UUID fstab entry, and mount at /labextra. Verify with df -h and sudo lvs.
• Add 512 MiB of swap non-destructively: either a new swap partition on free disk space or a new LVM swap volume. Activate it and confirm both old and new swap are active with swapon --show. Add the persistent fstab entry.
• Run findmnt --verify to confirm all fstab entries are valid. Run sudo mount -a and sudo swapon -a to ensure all entries are activated.
• Reboot the system. After rebooting, run df -h, swapon --show, and check that all new mounts are present and the original data is still intact.

Next: Getting Help from Local Documentation