docs: fix Crucible guide - single node works fine

Remove incorrect "3 downstairs minimum" requirement.
Single node Crucible deployment works and is in production.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Author: web-flow

docs/ma90-crucible-deployment-complete-guide.md

@@ -25,12 +25,6 @@ ssh root@still-fawn.maas "ls -la /tmp/crucible/target/release/crucible-downstair
 
 ## 🚨 CRITICAL LIMITATIONS
 
-### **Single-Sled Testing Not Supported**
-**CRUCIBLE REQUIRES MINIMUM 3 DOWNSTAIRS FOR TESTING** - Cannot test with single MA90 sled!
-- ❌ **1 downstairs**: Crucible upstairs/NBD fails to start
-- ✅ **3 downstairs**: Required for any Crucible testing (can be on same sled)
-- 📋 **Minimum deployment**: 3 MA90 sleds OR 3 downstairs processes on single sled
-
 ### **Performance Requirements**
 **NEVER USE 512-BYTE BLOCKS** - Performance is 10x worse than 4K blocks!
 - ❌ **512B blocks**: 6-11 MB/s  
@@ -416,36 +410,15 @@ ssh -i ~/.ssh/id_ed25519_pve ubuntu@proper-raptor.maas "ps aux | grep crucible"
 
 ## Next Steps
 
-### 💡 **BUDGET-FRIENDLY TESTING STRATEGY**
-With only 1 downstairs process, Crucible requires 3 total downstairs for functionality:
-- Crucible upstairs uses 3-way quorum for consensus-based operations
-- NBD server needs minimum 3 downstairs targets to start
-- **Solution**: Add 2 more downstairs processes (same sled = $30 testing vs 3 sleds = $90)
-
-### **Required Next Steps for Testing:**
-
-**Option A: Deploy 2 Additional MA90 Sleds (Production Fault Tolerance)**
-1. **Deploy 2 more MA90 sleds** using this exact process
-2. **Configure each with 1 downstairs** on port 3810
-3. **Test true 3-way distributed replication** across 3 separate sleds
-4. **Benefits**: Hardware fault tolerance, geographic distribution, 3x performance scaling
-
-**Option B: Add 2 More Downstairs to Current Sled (Cost-Effective Testing)**  
-
-**💡 SMART BUDGET APPROACH - Test full Crucible functionality for $30**
-- **Perfect for Learning**: Understand all Crucible features without expensive hardware
-- **Complete Testing**: NBD integration, replication, upstairs/downstairs communication
-- **Development Platform**: Ideal for experimentation and proof-of-concept work
-- **Upgrade Ready**: Easy migration to 3-sled production setup when needed
-
-1. **Create 2 additional regions** on proper-raptor (see NBD guide)
-2. **Start downstairs on ports 3811, 3812**
-3. **Test complete 3-downstairs functionality** on single sled
+### **After Deployment:**
+1. **Test NBD integration** with Proxmox VMs (see `docs/crucible-proxmox-nbd-integration-guide.md`)
+2. **Set up systemd services** for persistent operation
+3. **Monitor performance** and storage health
 
-### **After 3 Downstairs Available:**
-4. **Test NBD integration** with Proxmox VMs (see `docs/crucible-proxmox-nbd-integration-guide.md`)
-5. **Set up systemd services** for persistent operation
-6. **Monitor performance** and storage health
+### **Optional: Multi-Sled for Fault Tolerance**
+For production fault tolerance, deploy additional MA90 sleds:
+- Each sled runs 1 downstairs on port 3810
+- Benefits: Hardware fault tolerance, geographic distribution
 
 ---