Proxmox Cluster Auto-Update Script

📋 Overview

This Bash script automates the update process for all nodes in a Proxmox cluster. It sequentially performs a complete system update on each node and sends an email notification with the status.

✨ Features

• ✅ Automatic sequential updates of all cluster nodes
• ✅ Automatic SSH key setup between nodes (optional with password)
• ✅ Lock mechanism prevents concurrent executions
• ✅ Automatic installation of needrestart if not present
• ✅ Complete system update: apt update, full-upgrade, autoremove, clean
• ✅ Reboot status check using two methods:
  • needs-restarting -r (needrestart package)
  • /var/run/reboot-required (Debian standard)
• ✅ Detailed logging of all actions
• ✅ Email notification with summary
• ✅ No automatic reboot - administrator maintains manual control

📦 Prerequisites

On the coordinating node:

apt-get install -y flock mailutils

Optional (for automatic SSH key setup with password):

apt-get install -y sshpass

All nodes:

• Proxmox VE installed (Postfix is already included)
• Root SSH access between nodes
• Debian-based system (tested with Debian 13)

🚀 Installation

1. Download the script

cd /root
git clone https://github.com/MrMasterbay/proxmox-auto-update.git
chmod +x proxmox-auto-update.sh

Or create manually:

nano /root/proxmox-auto-update.sh
# Paste script content
chmod +x /root/proxmox-auto-update.sh

2. Adjust configuration

Open the script and modify the variables in the configuration section:

nano /root/proxmox-auto-update.sh

Important configuration options:

# Email recipient
MAIL_TO="root@localhost"              # Change to your email

# SSH password (optional - only if SSH keys are not yet configured)
SSH_PASSWORD=""                        # Leave empty or enter password

# Manual node list (only if no cluster is configured)
MANUAL_NODES=""                        # Example: "node2.domain.com node3.domain.com"

3. Set up SSH keys (if not already configured)

Option A: Automatic setup with password

Set in the script:

SSH_PASSWORD="your_root_password"

The script will automatically set up SSH keys on first run.

Option B: Manual setup

# Generate SSH key (if not present)
ssh-keygen -t ed25519 -f /root/.ssh/id_ed25519 -N ""

# Copy key to all other nodes
ssh-copy-id root@node2.domain.com
ssh-copy-id root@node3.domain.com

4. Test

Run the script manually:

/root/proxmox-auto-update.sh

Check the log:

cat /var/log/proxmox-cluster-auto-update.log

⏰ Automation with Cron

Daily updates at 3:00 AM

crontab -e

Add:

0 3 * * * /root/proxmox-auto-update.sh

Weekly updates (Sunday, 3:00 AM)

0 3 * * 0 /root/proxmox-auto-update.sh

Monthly updates (1st of month, 3:00 AM)

0 3 1 * * /root/proxmox-auto-update.sh

📧 Email Configuration

Postfix is pre-installed on Proxmox

Proxmox VE comes with Postfix pre-installed and configured. You only need to verify the configuration.

Verify Postfix configuration

# Check Postfix status
systemctl status postfix

# View main configuration
postconf -n

# Test email delivery
echo "Test message" | mail -s "Test Mail" your@email.com

Configure Postfix for external delivery

Edit the main configuration:

nano /etc/postfix/main.cf

For internet delivery, ensure these settings:

myhostname = node1.yourdomain.com
mydestination = $myhostname, localhost.$mydomain, localhost
relayhost = 
inet_interfaces = all

Restart Postfix:

systemctl restart postfix

For external SMTP relay (e.g., Gmail, Office365)

If your ISP blocks port 25, use an SMTP relay:

nano /etc/postfix/main.cf

Add:

relayhost = [smtp.gmail.com]:587
smtp_use_tls = yes
smtp_sasl_auth_enable = yes
smtp_sasl_security_options = noanonymous
smtp_sasl_password_maps = hash:/etc/postfix/sasl_passwd
smtp_tls_CAfile = /etc/ssl/certs/ca-certificates.crt

Create credentials file:

nano /etc/postfix/sasl_passwd

Content:

[smtp.gmail.com]:587 your@gmail.com:your-app-password

Secure and activate:

chmod 600 /etc/postfix/sasl_passwd
postmap /etc/postfix/sasl_passwd
systemctl restart postfix

Test email delivery

echo "Test from Proxmox" | mail -s "Test Subject" your@email.com

# Check mail logs
tail -f /var/log/mail.log

# Check mail queue
mailq

📊 Log Files

Main log file

/var/log/proxmox-cluster-auto-update.log

Contains:

• Timestamps of all actions
• Update progress per node
• Errors and warnings
• Reboot status
• Summary

View logs

# Full log
cat /var/log/proxmox-cluster-auto-update.log

# Last 50 lines
tail -n 50 /var/log/proxmox-cluster-auto-update.log

# Live view while script is running
tail -f /var/log/proxmox-cluster-auto-update.log

# Search for errors
grep -i error /var/log/proxmox-cluster-auto-update.log

Set up log rotation

nano /etc/logrotate.d/proxmox-auto-update

Content:

/var/log/proxmox-cluster-auto-update.log {
    weekly
    rotate 12
    compress
    delaycompress
    missingok
    notifempty
}

🔍 Cluster Detection

The script automatically detects cluster nodes using three methods:

1. Proxmox Cluster Manager (pvecm)

Preferred method for Proxmox clusters:

pvecm nodes

2. /etc/pve/nodes directory

Fallback method:

ls /etc/pve/nodes

3. Manual node list

For non-cluster setups or as override:

MANUAL_NODES="node2.domain.com node3.domain.com"

📝 Example Output

Successful execution

====================================================================
Proxmox Cluster Auto-Update started: 2025-10-05 03:00:01
Coordinating Node: node1.domain.com
====================================================================

[03:00:01] Updating LOCAL node: node1.domain.com
[03:00:01] apt update …
[03:00:03] apt full-upgrade -y …
[03:00:25] apt autoremove -y …
[03:00:26] apt clean …
[03:00:26] Local node update completed
⚠️  /var/run/reboot-required exists → reboot required

[03:00:27] Found cluster nodes: node2.domain.com node3.domain.com
[03:00:27] Starting updates on remote nodes...

====================================================================
[03:00:27] Updating REMOTE node: node2.domain.com
====================================================================
[03:00:28] ========== Update starting on node2.domain.com ==========
[03:00:45] ========== Update completed on node2.domain.com ==========
✅  /var/run/reboot-required absent → no reboot needed

====================================================================
[03:01:15] Updating REMOTE node: node3.domain.com
====================================================================
[03:01:16] ========== Update starting on node3.domain.com ==========
[03:01:33] ========== Update completed on node3.domain.com ==========
⚠️  /var/run/reboot-required exists → reboot required

--------------------------------------------------------------------
Proxmox Cluster Auto-Update finished: 2025-10-05 03:01:35

Cluster Update Summary:

• node1.domain.com: SUCCESS - Reboot required
• node2.domain.com: SUCCESS - No reboot needed
• node3.domain.com: SUCCESS - Reboot required
====================================================================

Email notification

Subject: [PROXMOX CLUSTER] Auto-Update completed on all nodes

Content:

The scheduled Proxmox cluster update finished on all nodes.

Cluster Update Summary:

• node1.domain.com: SUCCESS - Reboot required
• node2.domain.com: SUCCESS - No reboot needed
• node3.domain.com: SUCCESS - Reboot required

For full details, see the log file at /var/log/proxmox-cluster-auto-update.log on node1.domain.com.

🔧 Troubleshooting

Problem: SSH connection fails

Solution 1: Set up SSH keys manually

ssh-copy-id root@node2.domain.com

Solution 2: Set SSH_PASSWORD in script

SSH_PASSWORD="your_password"

Solution 3: Test SSH connection

ssh -v root@node2.domain.com

Problem: Script already running (lock file)

Symptom: Error message "another instance is already running"

Solution:

# Check if process is still running
ps aux | grep proxmox-cluster-auto-update

# If no process is running, remove lock file
rm /var/run/proxmox-auto-update.lock

Problem: No email received

Solution 1: Test mail configuration

echo "Test" | mail -s "Test" your@email.com
tail -f /var/log/mail.log

Solution 2: Check Postfix status

systemctl status postfix
journalctl -u postfix -n 50

Solution 3: Check mail queue

mailq
postqueue -p

Solution 4: Check if emails are being blocked

# Check if port 25 is blocked by ISP
telnet smtp.gmail.com 25

# If blocked, configure SMTP relay (see Email Configuration section)

Problem: Node not detected

Symptom: "No cluster nodes detected"

Solution 1: Check cluster status

pvecm status
pvecm nodes

Solution 2: Use manual node list

# Set in script:
MANUAL_NODES="node2.domain.com node3.domain.com"

Problem: Update fails

Symptom: "apt full-upgrade" error

Solution 1: Check repository issues

apt-get update
cat /etc/apt/sources.list

Solution 2: Check disk space

df -h
apt-get clean
apt-get autoremove

Solution 3: Check held packages

apt-mark showhold

🔒 Security Considerations

1. SSH Keys: Use SSH keys instead of passwords when possible
2. Password in script: If SSH_PASSWORD is set, protect the script:

   chmod 700 /root/proxmox-auto-update.sh

3. Log files: Do not contain passwords, but do contain system information
4. Root access: Script must run as root
5. Backups: Make backups before major updates

🎯 Best Practices

1. Schedule maintenance windows

# Cron for Sunday 3:00 AM (low load)
0 3 * * 0 /root/proxmox-auto-update.sh

2. Set up monitoring

# Nagios/Icinga check
/usr/lib/nagios/plugins/check_file_age -f /var/log/proxmox-cluster-auto-update.log -w 86400 -c 172800

3. Test environment

Test updates in a test environment before applying to production.

4. Staged updates

For large clusters: Update nodes in groups

# Group 1: Mondays
MANUAL_NODES="node2 node3"

# Group 2: Wednesdays
MANUAL_NODES="node4 node5"

5. Backup before updates

# Proxmox backup before cron job
0 2 * * 0 vzdump --all --storage backup-storage
0 3 * * 0 /root/proxmox-auto-update.sh

📚 Advanced Configuration

Update only specific nodes

# In script:
MANUAL_NODES="node2.domain.com node3.domain.com"

# Ignore nodes from pvecm - only use MANUAL_NODES
# Comment out pvecm detection (lines 165-190)

Different update times per node

Create multiple script variants:

# Script 1: Only nodes 1+2
cp proxmox-auto-update.sh update-group1.sh
# Set: MANUAL_NODES="node1 node2"

# Script 2: Only nodes 3+4
cp proxmox-auto-update.sh update-group2.sh
# Set: MANUAL_NODES="node3 node4"

# Crontab:
0 3 * * 1 /root/update-group1.sh  # Monday
0 3 * * 3 /root/update-group2.sh  # Wednesday

Custom APT options

Modify in script:

# Change line 31:
APTCMD="/usr/bin/apt-get -o Dpkg::Options::='--force-confold'"

Pre/Post-update hooks

Add custom commands:

# Before update (line 330):
echo "[$(date '+%H:%M:%S')] Running pre-update hook …"
/root/pre-update-hook.sh

# After update (line 375):
echo "[$(date '+%H:%M:%S')] Running post-update hook …"
/root/post-update-hook.sh

📬 Email Troubleshooting Guide

Check if Postfix is running

systemctl status postfix
systemctl enable postfix
systemctl start postfix

Test local mail delivery

echo "Test local" | mail -s "Test" root
cat /var/mail/root

Check Postfix configuration

postconf -n | grep -E 'myhostname|mydestination|relayhost'

View mail logs in real-time

tail -f /var/log/mail.log

Check mail queue and errors

mailq
postqueue -p
postcat -q QUEUE_ID  # Replace QUEUE_ID with actual ID from mailq

Flush mail queue

postqueue -f

Common Postfix fixes

Fix 1: Set proper hostname

hostnamectl set-hostname node1.yourdomain.com
postconf -e "myhostname = node1.yourdomain.com"
systemctl restart postfix

Fix 2: Allow localhost relay

postconf -e "mynetworks = 127.0.0.0/8 [::ffff:127.0.0.0]/104 [::1]/128"
systemctl restart postfix

Fix 3: Fix email aliases

echo "root: your@email.com" >> /etc/aliases
newaliases
systemctl restart postfix

🆘 Support & Contact

• Proxmox Forum: https://forum.proxmox.com
• Proxmox Documentation: https://pve.proxmox.com/wiki/Main_Page
• Debian Wiki: https://wiki.debian.org

📄 License

This script is freely available and can be modified and used as desired.

🔄 Changelog

Version 1.0 (2025-10-05)

• Initial version
• Automatic cluster node detection
• SSH key management
• Email notifications
• Detailed logging
• Reboot status detection

---

⚠️ IMPORTANT: This script does NOT perform automatic reboots. After updates requiring a reboot, you must manually restart the nodes!

# Check reboot status
cat /var/run/reboot-required

# Restart node
reboot