openboatmobile-ai/HERMES_VERIFICATION_CHECKLIST.md

Quick Reference: Hermes Deployment Status Check

For Current Deployment (Before Fixes)

If you're still SSH'd into the server from your initial deployment, run these checks:

Check 1: Is the systemd service running?

systemctl status hermes.service

Expected (BROKEN - before fix): Shows failed or inactive

Check 2: Does the Docker container exist?

docker ps -a | grep hermes

Expected (BROKEN - before fix): Container doesn't exist OR shows Exited status

Check 3: Check systemd journal for errors

journalctl -u hermes.service | tail -50

Expected (BROKEN - before fix): Error like "docker: command not found" or "file not found"

Check 4: Watch docker logs

docker logs hermes 2>&1 | head -20

Expected (BROKEN - before fix): Either no container, or errors about missing files

Check 5: Is Discord bot online?

# Go to Discord and check your server
# Look for the bot in members list

Expected (BROKEN - before fix): Shows Offline or doesn't appear

---

After Redeploying with Fixes

Run these verification commands immediately after deployment:

Quick Verification (< 1 minute)

# 1. Check service status
systemctl status hermes.service

# 2. Check Docker container
docker ps | grep hermes

# 3. Check port is listening
netstat -tlnp | grep 18789

Expected (FIXED):

• Service shows active (running)
• Container shows UP status
• Port 18789 shows LISTEN

Comprehensive Health Check (< 5 minutes)

/usr/local/bin/hermes-health-check.sh

Expected (FIXED): All checks show ✓

Detailed Logs

# Check what's happening in the container
docker logs -f hermes

# Use Ctrl+C to exit after 10-20 lines

Expected (FIXED):

[INFO] Hermes Agent Framework starting...
[INFO] Initializing gateway on port 18789
[INFO] Discord bot initialized

Discord Connectivity Test

# In your Discord server, type:
@hermes help

# Bot should respond within 5 seconds

Expected (FIXED): Bot is online and responds

---

Troubleshooting Matrix

Symptom	Check	Fix
Service shows failed	journalctl -u hermes.service	Redeploy with fixed template
Container Exited	docker logs hermes	Check the logs for errors
Port not listening	docker ps	Container not running
Docker permission denied	Check User= in service	Should be root now
Bot shows offline	Check Discord bot token	Verify in .env file
No container at all	docker ps -a	Image wasn't pulled, redeploy

---

Command Reference

Systemd Service

# Check status
systemctl status hermes.service

# View logs (last 50 lines)
journalctl -u hermes.service -n 50

# View logs with timestamps
journalctl -u hermes.service -f --all

# Restart service
systemctl restart hermes.service

# Stop service
systemctl stop hermes.service

# Start service
systemctl start hermes.service

Docker

# List running containers
docker ps

# List all containers (including stopped)
docker ps -a

# View container logs
docker logs hermes

# Follow logs (live)
docker logs -f hermes

# Show last 100 lines
docker logs --tail=100 hermes

# Inspect container
docker inspect hermes

Files to Check

# Configuration files
cat ~/.hermes/.env
cat ~/.hermes/config.yaml
cat ~/docker-compose.yml

# Check permissions
ls -la ~/.hermes/

# Check if Hermes healthcheck script exists
ls -la /usr/local/bin/hermes-health-check.sh

---

Before vs After Comparison

BEFORE These Fixes:

❌ systemctl status hermes.service
   → inactive (dead)
   
❌ docker ps
   → (no container)
   
❌ journalctl -u hermes.service
   → cannot open: "/home/hermes/docker-compose.yml"
   
❌ Discord bot
   → OFFLINE

AFTER These Fixes:

✓ systemctl status hermes.service
  → active (running)
  
✓ docker ps
  → hermes container UP 2 minutes
  
✓ journalctl -u hermes.service
  → [INFO] Hermes Agent started successfully
  
✓ Discord bot
  → ONLINE ✓

---

When to Seek Help

If after redeployment you still have issues:

1. Check HERMES_DEBUGGING.md in docs/ for detailed troubleshooting
2. Read HERMES_AUDIT_REPORT.md for what was fixed
3. Run health check: /usr/local/bin/hermes-health-check.sh
4. Share logs: docker logs hermes output
5. Check config: Verify Discord token, server ID, user IDs in ~/.hermes/.env

---

Redeploy Command

To apply all fixes:

cd ~/openboatmobile

# Option 1: Clean slate (recommended)
terraform destroy -auto-approve
source .env && terraform init && terraform apply

# Option 2: Update in-place
source .env && terraform apply -auto-approve

Then verify with:

ssh hermes@<SERVER_IP>
/usr/local/bin/hermes-health-check.sh