Nexus 3.0 Hard Boot Recovery - Complete Guide

Last Updated: January 13, 2026
Status: TESTED - Successfully recovered from hard reboot with zero data loss

What Happened

On January 12, 2026, the Nexus server experienced a hard reboot due to a locker Python script crash. When the server came back up:

• All Docker containers were running
• All data was intact on NVMe storage
• BUT: MCP tools couldn't connect to any environment
• Error: "invalid username-password pair or user is disabled"

Root Causes

1. Password Mismatch: Docker containers reverted to old weak passwords (*3Vault2025), while Locker stored strong alphanumeric passwords (tg6ap5, VFLtfX)

2. Port Mapping Issue: Docker containers ran Redis internally but ports weren't exposed to host after reboot. Fix: docker restart made ports listen

3. Replication Not Configured: Vault and operational containers weren't on the same Docker network, so master-slave replication couldn't work

Recovery Results

Environments Recovered: - KB: 687 keys, replication working - Workflow: 214 keys, replication working
- Locker: 55 keys (vault-only) - Total: 11,159+ keys across all environments - Data Loss: ZERO

See child articles for detailed recovery procedures.