test: add end-to-end agent deployment lifecycle test

tests/test_e2e_flow.py

@@ -0,0 +1,104 @@
+import pytest
+import subprocess
+import os
+import uuid
+import testinfra
+import time
+
+# Use the bin/infra wrapper for testing
+CLI_BIN = os.path.abspath(os.path.join(os.path.dirname(__file__), "..", "bin", "infra"))
+CONFIG_PATH = os.path.abspath(os.path.join(os.path.dirname(__file__), "..", "config.yaml"))
+SSH_KEY = os.path.expanduser("~/.ssh/id_ed25519_no_pass")
+ROUTER_PASS = "kpvoh58zhq2sq6ms"
+
+def run_infra(cmd, env=None):
+    full_env = os.environ.copy()
+    if env:
+        full_env.update(env)
+    full_env["ROUTER_PASS"] = ROUTER_PASS
+    full_cmd = [CLI_BIN, "--config", CONFIG_PATH] + cmd
+    return subprocess.run(full_cmd, capture_output=True, text=True, env=full_env)
+
+def get_testinfra_host(host_str):
+    return testinfra.get_host(
+        f"ssh://root@{host_str}",
+        ssh_identity_file=SSH_KEY,
+        ssh_extra_args="-o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null"
+    )
+
+@pytest.fixture
+def agent_context():
+    """Provides a unique context for an agent service deployment"""
+    uid = str(uuid.uuid4())[:8]
+    return {
+        "id": uid,
+        "name": f"agent-svc-{uid}",
+        "domain": f"agent-{uid}.loopaware.com",
+        "mac": f"aa:bb:cc:ee:ff:{uid[:2]}",
+        "port": 8080
+    }
+
+def test_e2e_agent_deployment_lifecycle(agent_context):
+    """
+    Tests the complete flow:
+    IPAM -> DB -> DNS -> Ingress -> Cloudflare -> Decommission
+    """
+    ctx = agent_context
+    print(f"\n[E2E] Starting deployment for {ctx['name']}")
+
+    # 1. IPAM: Find next free IP
+    res = run_infra(["ip", "next-free"])
+    assert res.returncode == 0
+    ip = res.stdout.strip()
+    print(f"[E2E] Allocated IP: {ip}")
+
+    # 2. Database: Provision PostgreSQL
+    res = run_infra(["db", "provision", ctx['name']])
+    assert res.returncode == 0
+    assert ctx['name'].replace("-", "_") in res.stdout
+    print("[E2E] Database provisioned")
+
+    # 3. DNS: Register DHCP Host
+    res = run_infra(["dns", "add-host", ctx['mac'], ip, ctx['name']])
+    assert res.returncode == 0
+    print("[E2E] DNS/DHCP host registered")
+
+    # 4. Ingress: Configure HAProxy
+    res = run_infra(["ingress", "add", ctx['domain'], ip, str(ctx['port'])])
+    assert res.returncode == 0
+    print(f"[E2E] Ingress configured for {ctx['domain']}")
+
+    # 5. Cloudflare: Add to DDNS list
+    res = run_infra(["cloudflare", "add-ddns", ctx['domain']])
+    assert res.returncode == 0
+    print("[E2E] Cloudflare DDNS registered")
+
+    # --- VERIFICATION PHASE ---
+    print("[E2E] Verifying infrastructure state...")
+    
+    # Verify HAProxy points to the CORRECT allocated IP
+    pve = get_testinfra_host("10.32.2.1")
+    # Wrap in single quotes to protect from local shell, double for remote
+    be_name = f"be_{ctx['domain'].replace('.', '_')}"
+    res = pve.run(f"pct exec 11220 -- grep -A 5 '{be_name}' /etc/haproxy/conf.d/99-dynamic.cfg")
+    assert res.rc == 0
+    assert ip in res.stdout
+    assert str(ctx['port']) in res.stdout
+
+    # Verify Database exists
+    db_host = get_testinfra_host("10.32.70.54")
+    db_name = ctx['name'].lower().replace("-", "_")
+    # Safer check: use psql -t -c to get raw list
+    res = db_host.run(f"su - postgres -c \"psql -t -c '\\l' | grep -qw {db_name}\"")
+    assert res.rc == 0
+
+    # --- CLEANUP / DECOMMISSION PHASE ---
+    print("[E2E] Starting decommissioning...")
+    res = run_infra(["decommission", "--domain", ctx['domain'], "--mac", ctx['mac']])
+    assert res.returncode == 0
+    
+    # Final state check: should be gone in sites.json
+    res = pve.run(f"pct exec 11220 -- grep '{ctx['domain']}' /etc/haproxy/dynamic/sites.json")
+    assert res.rc != 0
+    
+    print("[E2E] Flow completed successfully.")