Storage Management with LVM and Stratis on CentOS/RHEL

Effective storage management is crucial for enterprise environments. This guide covers Logical Volume Manager (LVM) and Stratis, providing flexible and scalable storage solutions for CentOS/RHEL systems.

LVM Fundamentals

LVM Architecture Overview

Understanding LVM components:

# LVM hierarchy:
# Physical Volumes (PV) -> Volume Groups (VG) -> Logical Volumes (LV)

# Check LVM version
lvm version

# Display LVM configuration
lvmconfig --typeconfig default

# View current LVM setup
pvs    # Physical volumes
vgs    # Volume groups
lvs    # Logical volumes

# Detailed view
pvdisplay
vgdisplay
lvdisplay

Creating LVM Storage

Set up basic LVM configuration:

# Prepare disks for LVM
parted /dev/sdb mklabel gpt
parted /dev/sdb mkpart primary 1MiB 100%
parted /dev/sdb set 1 lvm on

# Create physical volumes
pvcreate /dev/sdb1 /dev/sdc1 /dev/sdd1

# Create volume group
vgcreate vg_data /dev/sdb1 /dev/sdc1

# Create logical volumes
lvcreate -L 100G -n lv_database vg_data
lvcreate -L 50G -n lv_logs vg_data
lvcreate -l 100%FREE -n lv_backup vg_data

# Create filesystems
mkfs.xfs /dev/vg_data/lv_database
mkfs.ext4 /dev/vg_data/lv_logs
mkfs.xfs /dev/vg_data/lv_backup

# Mount logical volumes
mkdir -p /data/{database,logs,backup}
mount /dev/vg_data/lv_database /data/database
mount /dev/vg_data/lv_logs /data/logs
mount /dev/vg_data/lv_backup /data/backup

# Add to fstab
cat >> /etc/fstab <<EOF
/dev/vg_data/lv_database /data/database xfs defaults 0 0
/dev/vg_data/lv_logs /data/logs ext4 defaults 0 0
/dev/vg_data/lv_backup /data/backup xfs defaults 0 0
EOF

Advanced LVM Features

Volume Expansion and Reduction

Resize logical volumes dynamically:

# Extend volume group
pvcreate /dev/sdd1
vgextend vg_data /dev/sdd1

# Extend logical volume
lvextend -L +50G /dev/vg_data/lv_database
# Or use percentage
lvextend -l +50%FREE /dev/vg_data/lv_logs

# Resize filesystem
# XFS (can only grow)
xfs_growfs /data/database

# EXT4 (can grow and shrink)
resize2fs /dev/vg_data/lv_logs

# Reduce logical volume (EXT4 only)
umount /data/logs
e2fsck -f /dev/vg_data/lv_logs
resize2fs /dev/vg_data/lv_logs 40G
lvreduce -L 40G /dev/vg_data/lv_logs
mount /dev/vg_data/lv_logs /data/logs

LVM Snapshots

Create and manage snapshots:

# Create snapshot
lvcreate -L 10G -s -n lv_database_snap /dev/vg_data/lv_database

# Mount snapshot
mkdir /mnt/database_snap
mount -o ro /dev/vg_data/lv_database_snap /mnt/database_snap

# Monitor snapshot usage
lvs -o +snap_percent

# Merge snapshot (rollback)
umount /data/database
lvconvert --merge /dev/vg_data/lv_database_snap

# Remove snapshot
umount /mnt/database_snap
lvremove /dev/vg_data/lv_database_snap

Thin Provisioning

Implement thin provisioned volumes:

# Create thin pool
lvcreate -L 500G --thinpool tp_data vg_data

# Configure thin pool
lvchange --metadataprofile thin_performance tp_data

# Create thin volumes
lvcreate -V 200G --thin -n lv_thin_web vg_data/tp_data
lvcreate -V 300G --thin -n lv_thin_app vg_data/tp_data
lvcreate -V 400G --thin -n lv_thin_db vg_data/tp_data

# Monitor thin pool usage
lvs -o +thin_count,thin_data_percent,thin_metadata_percent

# Set thin pool thresholds
lvchange --monitor y --threshold 80 vg_data/tp_data

LVM RAID

Creating RAID Volumes

Configure software RAID with LVM:

# Create RAID 1 (mirror)
lvcreate --type raid1 -m 1 -L 100G -n lv_raid1 vg_data

# Create RAID 5
lvcreate --type raid5 -i 3 -L 300G -n lv_raid5 vg_data

# Create RAID 10
lvcreate --type raid10 -m 1 -i 2 -L 200G -n lv_raid10 vg_data

# Convert linear to RAID1
lvconvert --type raid1 -m 1 vg_data/lv_database

# Monitor RAID status
lvs -a -o +raid_sync_percent,raid_mismatch_count

RAID Management

Maintain RAID volumes:

# Replace failed device
lvconvert --repair vg_data/lv_raid1

# Add RAID images
lvconvert -m +1 vg_data/lv_raid1

# Remove RAID images
lvconvert -m -1 vg_data/lv_raid1

# Scrub RAID array
lvchange --syncaction check vg_data/lv_raid5
lvchange --syncaction repair vg_data/lv_raid5

# RAID reshape
lvconvert --stripes 4 vg_data/lv_raid5

LVM Cache and Performance

SSD Caching

Implement cache volumes:

# Create cache pool on SSD
pvcreate /dev/nvme0n1p1
vgextend vg_data /dev/nvme0n1p1

# Create cache data and metadata LVs
lvcreate -L 100G -n lv_cache_data vg_data /dev/nvme0n1p1
lvcreate -L 1G -n lv_cache_meta vg_data /dev/nvme0n1p1

# Create cache pool
lvconvert --type cache-pool --poolmetadata vg_data/lv_cache_meta \
    vg_data/lv_cache_data

# Attach cache to existing LV
lvconvert --type cache --cachepool vg_data/lv_cache_data \
    vg_data/lv_database

# Configure cache mode
lvchange --cachemode writethrough vg_data/lv_database
lvchange --cachemode writeback vg_data/lv_database

# Monitor cache performance
dmsetup status vg_data-lv_database
lvs -o +cache_used_blocks,cache_dirty_blocks

Performance Optimization

Tune LVM performance:

# Set read-ahead
lvchange -r 256 /dev/vg_data/lv_database

# Configure I/O scheduler
echo noop > /sys/block/dm-0/queue/scheduler

# Stripe volumes for performance
lvcreate -i 4 -I 64 -L 400G -n lv_striped vg_data

# Disable barriers for performance (risky)
mount -o nobarrier /dev/vg_data/lv_database /data/database

Stratis Storage Management

Stratis Introduction

Modern storage management with Stratis:

# Install Stratis
yum install stratisd stratis-cli

# Enable and start Stratis
systemctl enable --now stratisd

# Check Stratis version
stratis --version

# View Stratis daemon status
systemctl status stratisd

Creating Stratis Pools

Set up Stratis storage:

# Create pool
stratis pool create pool1 /dev/sdb /dev/sdc

# Add devices to pool
stratis pool add-data pool1 /dev/sdd

# List pools
stratis pool list

# Show pool details
stratis pool list pool1

# Create filesystems
stratis filesystem create pool1 fs_web
stratis filesystem create pool1 fs_database
stratis filesystem create pool1 fs_logs

# List filesystems
stratis filesystem list

# Mount Stratis filesystems
mkdir -p /stratis/{web,database,logs}
mount /stratis/pool1/fs_web /stratis/web
mount /stratis/pool1/fs_database /stratis/database
mount /stratis/pool1/fs_logs /stratis/logs

Stratis Snapshots

Manage filesystem snapshots:

# Create snapshot
stratis filesystem snapshot pool1 fs_database fs_database_snap

# List snapshots
stratis filesystem list pool1

# Mount snapshot
mkdir /mnt/database_snap
mount /stratis/pool1/fs_database_snap /mnt/database_snap

# Restore from snapshot
umount /stratis/database
stratis filesystem destroy pool1 fs_database
stratis filesystem snapshot pool1 fs_database_snap fs_database
mount /stratis/pool1/fs_database /stratis/database

# Remove snapshot
stratis filesystem destroy pool1 fs_database_snap

Storage Security

Encryption with LVM

Implement encrypted volumes:

# Install cryptsetup
yum install cryptsetup

# Create encrypted physical volume
cryptsetup luksFormat /dev/sde1
cryptsetup luksOpen /dev/sde1 encrypted_pv

# Add to LVM
pvcreate /dev/mapper/encrypted_pv
vgextend vg_secure /dev/mapper/encrypted_pv

# Create encrypted logical volume
lvcreate -L 50G -n lv_sensitive vg_data
cryptsetup luksFormat /dev/vg_data/lv_sensitive
cryptsetup luksOpen /dev/vg_data/lv_sensitive sensitive_data

# Create filesystem
mkfs.xfs /dev/mapper/sensitive_data
mount /dev/mapper/sensitive_data /secure

# Auto-mount encrypted volumes
cat >> /etc/crypttab <<EOF
sensitive_data /dev/vg_data/lv_sensitive none
EOF

# Key file setup
dd if=/dev/urandom of=/root/keyfile bs=1024 count=4
chmod 0400 /root/keyfile
cryptsetup luksAddKey /dev/vg_data/lv_sensitive /root/keyfile

Stratis Encryption

Encrypted Stratis pools:

# Create encrypted pool
stratis pool create --key-desc my_key encrypted_pool /dev/sdf

# Set up kernel keyring
stratis key set --capture-key my_key

# Create encrypted filesystem
stratis filesystem create encrypted_pool secure_fs

# Mount encrypted filesystem
mount /stratis/encrypted_pool/secure_fs /secure

Disaster Recovery

LVM Backup and Restore

Backup LVM configuration:

# Backup LVM metadata
vgcfgbackup vg_data

# Backup all VG metadata
vgcfgbackup

# View backup files
ls -la /etc/lvm/backup/

# Restore VG metadata
vgcfgrestore vg_data

# Archive management
ls -la /etc/lvm/archive/
vgcfgrestore --list vg_data

# Full LVM backup script
cat > /usr/local/bin/lvm-backup.sh <<'EOF'
#!/bin/bash
BACKUP_DIR="/backup/lvm"
mkdir -p $BACKUP_DIR

# Backup all VG configurations
for vg in $(vgs --noheadings -o vg_name); do
    vgcfgbackup -f $BACKUP_DIR/${vg}_$(date +%Y%m%d).vg $vg
done

# Backup partition tables
for disk in $(lsblk -dpno NAME,TYPE | grep disk | awk '{print $1}'); do
    sfdisk -d $disk > $BACKUP_DIR/$(basename $disk)_$(date +%Y%m%d).part
done
EOF
chmod +x /usr/local/bin/lvm-backup.sh

Stratis Backup

Backup Stratis configurations:

# Export pool information
stratis pool list --output-format json > /backup/stratis-pools.json

# Filesystem information
stratis filesystem list --output-format json > /backup/stratis-fs.json

# Create backup script
cat > /usr/local/bin/stratis-backup.sh <<'EOF'
#!/bin/bash
BACKUP_DIR="/backup/stratis"
mkdir -p $BACKUP_DIR

# Snapshot all filesystems
for pool in $(stratis pool list | tail -n +2 | awk '{print $1}'); do
    for fs in $(stratis filesystem list $pool | tail -n +2 | awk '{print $1}'); do
        stratis filesystem snapshot $pool $fs ${fs}_backup
    done
done
EOF
chmod +x /usr/local/bin/stratis-backup.sh

Monitoring and Maintenance

Storage Monitoring

Monitor storage health:

# LVM monitoring script
cat > /usr/local/bin/storage-monitor.sh <<'EOF'
#!/bin/bash

echo "=== Storage Health Check ==="

# Check PV health
echo "Physical Volumes:"
pvs -o +pv_missing,pv_allocatable

# Check VG health
echo -e "\nVolume Groups:"
vgs -o +vg_missing_pv_count

# Check thin pool usage
echo -e "\nThin Pools:"
lvs --type thin-pool -o +data_percent,metadata_percent

# Check snapshot usage
echo -e "\nSnapshots:"
lvs --type snapshot -o +snap_percent

# Check RAID health
echo -e "\nRAID Volumes:"
lvs --type raid -o +raid_sync_percent,raid_health

# Check Stratis pools
if command -v stratis &> /dev/null; then
    echo -e "\nStratis Pools:"
    stratis pool list
fi

# Disk space alerts
df -h | awk '$5 > 80 {print "WARNING: " $6 " is " $5 " full"}'
EOF
chmod +x /usr/local/bin/storage-monitor.sh

Maintenance Tasks

Regular maintenance procedures:

# LVM maintenance
# Check and repair VG metadata
vgck vg_data

# Activate missing volumes
vgchange -ay --activationmode partial

# Remove missing PVs
vgreduce --removemissing vg_data

# Defragment thin pool
lvchange --discards passdown vg_data/tp_data
fstrim -v /data/thin_volume

# Stratis maintenance
# Refresh Stratis cache
systemctl restart stratisd

# Check pool integrity
stratis pool check pool1

Advanced Configurations

Multipath Storage

Configure multipath with LVM:

# Install multipath tools
yum install device-mapper-multipath

# Configure multipath
mpathconf --enable --with_multipathd y

# Create multipath LVM
pvcreate /dev/mapper/mpatha
vgcreate vg_multipath /dev/mapper/mpatha

# Multipath configuration
cat >> /etc/multipath.conf <<EOF
blacklist {
    devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st)[0-9]*"
}

defaults {
    user_friendly_names yes
    find_multipaths yes
    failback immediate
}
EOF

Cluster Shared Storage

LVM in cluster environments:

# Install cluster LVM
yum install lvm2-cluster

# Configure LVM for clustering
lvmconf --enable-cluster

# Create clustered VG
vgcreate -cy vg_cluster /dev/sdb1

# Enable cluster locking
systemctl enable --now lvm2-lvmlockd

Performance Tuning

I/O Optimization

Optimize storage performance:

# LVM stripe optimization
lvcreate -i 4 -I 256K -L 1T -n lv_performance vg_data

# Cache policy tuning
lvchange --cachepolicy smq vg_data/lv_cached
lvchange --cachesettings migration_threshold=16384 vg_data/lv_cached

# Read-ahead tuning
for lv in $(lvs --noheadings -o lv_path); do
    blockdev --setra 4096 $lv
done

# Disable atime updates
mount -o noatime,nodiratime /dev/vg_data/lv_database /data/database

Best Practices

Storage Design Guidelines

Plan for growth: Size VGs with expansion in mind
Use meaningful names: Follow naming conventions
Regular backups: Backup both data and metadata
Monitor usage: Set up alerts for capacity
Document layout: Maintain storage documentation
Test recovery: Regular DR drills

Health Check Script

#!/bin/bash
# Comprehensive storage health check

echo "=== Storage System Audit ==="
date

# Check for failed PVs
failed_pvs=$(pvs -o pv_name,vg_name,pv_attr | grep -c "m-")
if [ $failed_pvs -gt 0 ]; then
    echo "CRITICAL: $failed_pvs failed physical volumes detected"
fi

# Check VG capacity
vgs --noheadings -o vg_name,vg_free_percent | while read vg free; do
    free_int=${free%.*}
    if [ $free_int -lt 10 ]; then
        echo "WARNING: Volume group $vg has only $free% free space"
    fi
done

# Check thin pool usage
lvs --type thin-pool -o lv_name,data_percent --noheadings | while read pool usage; do
    usage_int=${usage%.*}
    if [ $usage_int -gt 80 ]; then
        echo "WARNING: Thin pool $pool is $usage% full"
    fi
done

# Check Stratis status
if systemctl is-active stratisd &>/dev/null; then
    stratis pool list | tail -n +2 | while read pool size used; do
        echo "Stratis pool $pool: Size=$size, Used=$used"
    done
fi

echo "Storage audit complete"

Conclusion

LVM and Stratis provide flexible storage management solutions for CentOS/RHEL systems. LVM offers mature, feature-rich volume management with extensive configurability, while Stratis provides a modern, simplified approach to storage management. Choose the appropriate solution based on your requirements and leverage their advanced features for optimal storage utilization and performance.

Storage Management with LVM and Stratis on CentOS/RHEL

Need Professional CentOS/RHEL Support?

Storage Management with LVM and Stratis on CentOS/RHEL

LVM Fundamentals

LVM Architecture Overview

Creating LVM Storage

Advanced LVM Features

Volume Expansion and Reduction

LVM Snapshots

Thin Provisioning

LVM RAID

Creating RAID Volumes

RAID Management

LVM Cache and Performance

SSD Caching

Performance Optimization

Stratis Storage Management

Stratis Introduction

Creating Stratis Pools

Stratis Snapshots

Storage Security

Encryption with LVM

Stratis Encryption

Disaster Recovery

LVM Backup and Restore

Stratis Backup

Monitoring and Maintenance

Storage Monitoring

Maintenance Tasks

Advanced Configurations

Multipath Storage

Cluster Shared Storage

Performance Tuning

I/O Optimization

Best Practices

Storage Design Guidelines

Health Check Script

Conclusion