proxmox/docs/archive/completion/PROXMOX_PVE_PVE2_FIX_COMPLETE.md

Proxmox VE Fix Complete - pve and pve2

Date: 2025-01-20
Status: ✅ ALL ISSUES RESOLVED

---

Issues Fixed

Root Cause

The primary issue was hostname resolution failure. The pve-cluster service could not resolve the hostname "pve" or "pve2" to a non-loopback IP address, causing:

• pve-cluster service to fail
• /etc/pve filesystem not mounting
• SSL certificates not accessible
• pveproxy workers crashing

Error Message

Unable to resolve node name 'pve' to a non-loopback IP address - missing entry in '/etc/hosts' or DNS?

---

Fixes Applied

1. Hostname Resolution Fix

Script: scripts/fix-proxmox-hostname-resolution.sh

What it did:

• Added proper entries to /etc/hosts on both hosts
• Ensured hostnames resolve to their actual IP addresses (not loopback)
• Added both current hostname (pve/pve2) and correct hostname (r630-01/r630-02)

Results:

• ✅ pve-cluster service started successfully on both hosts
• ✅ /etc/pve filesystem is now mounted
• ✅ SSL certificates are accessible

2. SSL and Cluster Service Fix

Script: scripts/fix-proxmox-ssl-cluster.sh

What it did:

• Regenerated SSL certificates
• Restarted all Proxmox services in correct order
• Verified service status

Results:

• ✅ All services running
• ✅ Web interface accessible (HTTP 200)
• ✅ No worker exit errors

---

Current Status

pve (192.168.11.11 - r630-01)

Service	Status	Notes
pve-cluster	✅ Active (running)	Cluster filesystem mounted
pvestatd	✅ Active (running)	Status daemon working
pvedaemon	✅ Active (running)	API daemon working
pveproxy	✅ Active (running)	Web interface accessible
Web Interface	✅ Accessible	HTTP Status: 200
Port 8006	✅ Listening	Workers running normally

pve2 (192.168.11.12 - r630-02)

Service	Status	Notes
pve-cluster	✅ Active (running)	Cluster filesystem mounted
pvestatd	✅ Active (running)	Status daemon working
pvedaemon	✅ Active (running)	API daemon working
pveproxy	✅ Active (running)	Web interface accessible
Web Interface	✅ Accessible	HTTP Status: 200
Port 8006	✅ Listening	Workers running normally

---

/etc/hosts Configuration

pve (192.168.11.11)

192.168.11.11    pve pve.sankofa.nexus r630-01 r630-01.sankofa.nexus

pve2 (192.168.11.12)

192.168.11.12    pve2 pve2.sankofa.nexus r630-02 r630-02.sankofa.nexus

Key Point: The hostname (pve/pve2) must resolve to the actual IP address (192.168.11.11/12), not to 127.0.0.1. This is required for pve-cluster to function.

---

Cluster Status

Both nodes are in a cluster:

• Cluster Name: h
• Config Version: 3
• Transport: knet
• Status: Operational

---

Verification

Web Interface Access

# pve
curl -k https://192.168.11.11:8006/
# Returns: HTTP 200 ✅

# pve2
curl -k https://192.168.11.12:8006/
# Returns: HTTP 200 ✅

Service Status

# Check services on pve
ssh root@192.168.11.11 "systemctl status pve-cluster pvestatd pvedaemon pveproxy"

# Check services on pve2
ssh root@192.168.11.12 "systemctl status pve-cluster pvestatd pvedaemon pveproxy"

No Worker Exits

# Check for worker exit errors
ssh root@192.168.11.11 "journalctl -u pveproxy -n 50 | grep 'worker exit'"
# Should return: No recent worker exit errors ✅

---

Scripts Created

1. scripts/diagnose-proxmox-hosts.sh

   • Comprehensive diagnostic tool
   • Tests connectivity, SSH, and all Proxmox services
   • Usage: ./scripts/diagnose-proxmox-hosts.sh [pve|pve2|both]

2. scripts/fix-proxmox-hostname-resolution.sh

   • Fixes hostname resolution issues
   • Updates /etc/hosts with correct entries
   • Usage: ./scripts/fix-proxmox-hostname-resolution.sh

3. scripts/fix-proxmox-ssl-cluster.sh

   • Fixes SSL and cluster service issues
   • Regenerates certificates and restarts services
   • Usage: ./scripts/fix-proxmox-ssl-cluster.sh [pve|pve2|both]

---

Lessons Learned

1. Hostname Resolution is Critical

   • Proxmox VE requires hostnames to resolve to non-loopback IPs
   • /etc/hosts must have proper entries
   • DNS alone may not be sufficient

2. Service Dependencies

   • pve-cluster must be running before other services
   • /etc/pve filesystem must be mounted for SSL certificates
   • Services must be started in correct order

3. Cluster Filesystem

   • pmxcfs (Proxmox Cluster File System) is required
   • It provides /etc/pve as a FUSE filesystem
   • Without it, SSL certificates and configuration are inaccessible

---

Next Steps

1. ✅ Monitor Services

   • Watch for any worker exit errors
   • Verify web interface remains accessible

2. Consider Hostname Migration

   • Current hostnames: pve, pve2
   • Correct hostnames: r630-01, r630-02
   • Migration can be done later if needed (see HOSTNAME_MIGRATION_GUIDE.md)

3. Document Cluster Configuration

   • Document cluster setup
   • Note any cluster-specific requirements

---

Related Documentation

• Proxmox Issues Analysis - Original issue analysis
• Hostname Migration Guide - How to change hostnames
• R630-04 Troubleshooting - Similar issues on r630-04

---

Last Updated: 2025-01-20
Status: ✅ All Issues Resolved
Both hosts are now fully operational!