Manage disk space (df, du, ncdu)

A full disk on a HolyCloud Linux VPS can block logs, databases, updates, and even SSH. Learn to measure, locate, and clean space with standard tools df, du, and ncdu.

Prerequisites

• HolyCloud VPS Ubuntu/Debian with sudo access
• Active SSH session
• Ideally a HolyCloud snapshot before large file deletions

Step 1: overview with df

Show usage per partition:

df -hT

Useful columns:

| Column | Meaning |

|---------|---------------|

| Type | ext4, xfs, tmpfs… |

| Size / Used / Avail | Size and free space |

| Use% | Usage percentage — alert if > 85% |

| Mounted on | Mount point (/, /var, additional volume) |

Inode space (many small files):

df -hi

On HolyCloud, the main disk is usually mounted on /. An additional volume may appear on /mnt/data or similar.

Step 2: find large directories with du

Top 20 folders at root (may take time):

sudo du -xh --max-depth=1 / 2>/dev/null | sort -hr | head -20

Target /var (logs, databases, apt cache):

sudo du -xh --max-depth=1 /var | sort -hr | head -15

System logs:

sudo du -sh /var/log/*
sudo journalctl --disk-usage

Step 3: interactive analysis with ncdu

Install ncdu:

sudo apt update
sudo apt install -y ncdu

Scan root:

sudo ncdu /

Navigation: arrow keys, Enter to enter, d to delete (careful), q to quit.

Scan only /var/log:

sudo ncdu /var/log

Step 4: common safe cleanup

APT package cache:

sudo apt autoremove -y
sudo apt clean

Old kernels (Ubuntu, keep the current one):

dpkg -l 'linux-image-*' | grep ^ii
sudo apt autoremove --purge -y

systemd journal (limit to 500 MB):

sudo journalctl --vacuum-size=500M

Compressed or rotated Nginx/Apache logs — verify before deletion:

sudo ls -lh /var/log/nginx/
sudo truncate -s 0 /var/log/nginx/access.log

Docker containers (if installed):

docker system df
docker system prune -a

Warning: do not delete /var/lib/mysql or /var/lib/postgresql without a backup.

Step 5: proactive monitoring

Weekly cron script (e.g. /usr/local/bin/check-disk.sh):

#!/bin/bash
THRESH=85
USE=$(df / | awk 'NR==2 {print $5}' | tr -d '%')
if [ "$USE" -ge "$THRESH" ]; then
  echo "ALERTE disque : ${USE}% utilisé sur $(hostname)" | logger -t disk-alert
fi

sudo chmod +x /usr/local/bin/check-disk.sh
echo '0 8 * * 1 root /usr/local/bin/check-disk.sh' | sudo tee /etc/cron.d/disk-check

In the HolyCloud customer area, also watch disk usage graphs if available.

Step 6: expand disk on HolyCloud side

If cleanup is not enough:

1. Increase disk size or migrate to a higher plan from the HolyCloud panel
2. After hypervisor resize, extend the Linux partition (GPT, growpart, resize2fs — see the additional volume doc)

Verification

df -hT
sudo du -sh /var/log /var/cache /tmp /home

Need help?

• No space left on device: free /tmp and /var/log first
• Disk full but few visible files: check docker, local snapshots, deleted files still open (lsof | grep deleted)
• HolyCloud support: df -hT output, ncdu or du on /var, VPS plan and contracted disk size