propagation: single registry drives all node references

New bin/netvm-registry.py parses NODES.md (node -> cdp_port); every consumer reads from it instead of hardcoding: - onboard-driver.py: CDP_PORTS dict -> registry lookup (new nodes work) - muse-signin.py: hardcoded 9410 -> --node/--cdp-port args - muse-chat-api.py: hardcoded ACCOUNTS -> registry-built - agent-health.sh: hardcoded muse/pip blocks -> loop over all active nodes (646 and opm now get health coverage too) NODES.md: record opm node (9440).
2026-10-03 21:28:36 +00:00
parent 1191cbd86e
commit a7639e2341
6 changed files with 156 additions and 36 deletions
@@ -11,3 +11,4 @@ Unified naming: node == agent == profile == API account.
 - 2026-10-03: Renamed smoke->muse, phone-test->pip for unified naming.
  Profiles preserved, sessions persisted (muse). WireGuard identities renamed.
 | 646 | warp-646 | 104.28.195.181 | 9430 | active | 646 (phone_otp, first Meta Account option) |
+| opm | warp-opm | 104.28.195.181 | 9440 | active | opm (yourfriendnico@proton.me, email_otp, Nico Parada) |
@@ -49,28 +49,27 @@ restart_browser() {
 # Main
 echo "=== Health check $(date -Iseconds) ===" >> "$LOG"

-# muse (node muse, CDP 9410)
-if ! check_agent "muse" "muse" "9410"; then
-    restart_browser "muse" "9410"
-    sleep 5
-    if ! check_agent "muse" "muse" "9410"; then
-        echo "$(date -Iseconds) muse: CRITICAL - still down after restart" >> "$LOG"
-        # TODO: Alert operator (e.g., via board post or email)
-    else
-        echo "$(date -Iseconds) muse: RECOVERED after restart" >> "$LOG"
+check_one() {
+    local agent=$1
+    local cdp_port=$2
+    # node == agent == profile (unified naming)
+    if ! check_agent "$agent" "$agent" "$cdp_port"; then
+        restart_browser "$agent" "$cdp_port"
+        sleep 5
+        if ! check_agent "$agent" "$agent" "$cdp_port"; then
+            echo "$(date -Iseconds) $agent: CRITICAL - still down after restart" >> "$LOG"
+            # TODO: Alert operator (e.g., via board post or email)
+        else
+            echo "$(date -Iseconds) $agent: RECOVERED after restart" >> "$LOG"
+        fi
    fi
-fi
+}

-# pip (node pip, CDP 9420)
-if ! check_agent "pip" "pip" "9420"; then
-    restart_browser "pip" "9420"
-    sleep 5
-    if ! check_agent "pip" "pip" "9420"; then
-        echo "$(date -Iseconds) pip: CRITICAL - still down after restart" >> "$LOG"
-    else
-        echo "$(date -Iseconds) pip: RECOVERED after restart" >> "$LOG"
-    fi
-fi
+# Every active node in the NODES.md registry gets checked — new nodes
+# propagate automatically, no per-node blocks to add.
+"$NETVM_BIN/netvm-registry.py" 2>/dev/null | while IFS=: read -r node port; do
+    [ -n "$node" ] && [ -n "$port" ] && check_one "$node" "$port"
+done

 # Trim log (keep last 1000 lines)
 tail -1000 "$LOG" > "$LOG.tmp" && mv "$LOG.tmp" "$LOG"
@@ -16,14 +16,24 @@ Usage:
  muse-chat-api.py --account <agent> wait [timeout]
  muse-chat-api.py --account <agent> approvals  # check pending approvals
 """
-import json, urllib.request, websocket, time, sys, argparse
+import json, urllib.request, websocket, time, sys, argparse, importlib.util

-ACCOUNTS = {
-    "muse": ("muse", "http://127.0.0.1:9410/json/list"),
-    "pip": ("pip", "http://127.0.0.1:9420/json/list"),
-    "646": ("646", "http://127.0.0.1:9430/json/list"),
-    "opm": ("opm", "http://127.0.0.1:9440/json/list"),
-}
+
+def _load_accounts():
+    """Build the accounts table from the NODES.md registry — new nodes
+    propagate automatically instead of being hardcoded here."""
+    path = "/home/super/Projects/NetVM/bin/netvm-registry.py"
+    spec = importlib.util.spec_from_file_location("netvm_registry", path)
+    mod = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(mod)
+    accounts = {}
+    for node, rec in mod.load().items():
+        accounts[node] = (node, "http://127.0.0.1:%d/json/list" %
+                          rec["cdp_port"])
+    return accounts
+
+
+ACCOUNTS = _load_accounts()

 # IPs we trust for auto-approval (our infrastructure)
 TRUSTED_IPS = {
@@ -16,12 +16,25 @@ The operator then obtains the OTP (via chat with user) and re-runs:
 This keeps credentials out of the automation — the OTP is provided
 transiently via the operator, never stored.
 """
-import json, urllib.request, websocket, time, sys, argparse
+import json, urllib.request, websocket, time, sys, argparse, importlib.util

-CDP_URL = "http://127.0.0.1:9410/json/list"
+CDP_PORT_DEFAULT = 9410

-def get_page():
-    with urllib.request.urlopen(CDP_URL, timeout=5) as r:
+
+def _registry_port(node):
+    try:
+        path = "/home/super/Projects/NetVM/bin/netvm-registry.py"
+        spec = importlib.util.spec_from_file_location("netvm_registry", path)
+        mod = importlib.util.module_from_spec(spec)
+        spec.loader.exec_module(mod)
+        return mod.port_for(node)
+    except Exception:
+        return None
+
+
+def get_page(port):
+    cdp_url = "http://127.0.0.1:%s/json/list" % port
+    with urllib.request.urlopen(cdp_url, timeout=5) as r:
        ts = json.load(r)
    pages = [t for t in ts if t.get('type') == 'page']
    if not pages:
@@ -41,9 +54,20 @@ def main():
    p = argparse.ArgumentParser()
    p.add_argument('--email', required=True)
    p.add_argument('--otp', default=None)
+    p.add_argument('--node', default=None,
+                   help='node name: CDP port comes from the NODES.md registry')
+    p.add_argument('--cdp-port', default=None,
+                   help='explicit CDP port (overrides --node lookup)')
    args = p.parse_args()

-    page = get_page()
+    if args.cdp_port:
+        port = args.cdp_port
+    elif args.node:
+        port = _registry_port(args.node) or CDP_PORT_DEFAULT
+    else:
+        port = CDP_PORT_DEFAULT
+
+    page = get_page(port)
    ws = websocket.create_connection(page['webSocketDebuggerUrl'], timeout=15)

    # Step 1: Check if already logged in
@@ -0,0 +1,77 @@
+#!/usr/bin/env python3
+"""
+netvm-registry.py — single source of truth for node -> CDP port mapping.
+
+Parses NODES.md (the fleet registry). Every script that needs a node's
+CDP port reads it from here instead of hardcoding — new nodes propagate
+automatically.
+
+Usage:
+    from netvm_registry import port_for, load
+    port_for("muse")          # -> 9410 (int), None if unknown
+
+CLI:
+    netvm-registry.py               # prints "node:port" lines for active nodes
+    netvm-registry.py <node>        # prints just the port (for bash)
+"""
+import os
+import re
+import sys
+
+REGISTRY = os.path.join(os.path.dirname(os.path.dirname(
+    os.path.abspath(__file__))), "NODES.md")
+
+
+def load(registry_path=None):
+    """Return {node: {netns, egress_ip, cdp_port, status, agent}}."""
+    path = registry_path or REGISTRY
+    nodes = {}
+    try:
+        with open(path) as f:
+            for line in f:
+                line = line.rstrip("\n")
+                if not line.startswith("|"):
+                    continue
+                cells = [c.strip() for c in line.strip("|").split("|")]
+                if len(cells) < 6:
+                    continue
+                node, netns, egress, port, status, agent = cells[:6]
+                if node in ("node", "") or not re.fullmatch(r"[a-z0-9-]+",
+                                                            node):
+                    continue
+                if node.startswith("-"):
+                    continue
+                try:
+                    port_n = int(port)
+                except ValueError:
+                    continue
+                nodes[node] = {"netns": netns, "egress_ip": egress,
+                               "cdp_port": port_n, "status": status,
+                               "agent": agent}
+    except FileNotFoundError:
+        pass
+    return nodes
+
+
+def port_for(node, registry_path=None):
+    """CDP port for a node, or None if the node isn't registered."""
+    rec = load(registry_path).get(node)
+    return rec["cdp_port"] if rec else None
+
+
+def active_nodes(registry_path=None):
+    """{node: port} for nodes with status == active."""
+    return {n: r["cdp_port"] for n, r in load(registry_path).items()
+            if r["status"] == "active"}
+
+
+if __name__ == "__main__":
+    if len(sys.argv) == 2:
+        port = port_for(sys.argv[1])
+        if port is None:
+            print("unknown node: %s" % sys.argv[1], file=sys.stderr)
+            sys.exit(1)
+        print(port)
+    else:
+        for node, port in sorted(active_nodes().items()):
+            print("%s:%d" % (node, port))
@@ -16,13 +16,21 @@ VM->bl SSH stdin pipe.
 Part of the cred onboarding module (front-door repo, docs/CRED-MODULE.md).
 """
 import argparse
+import importlib.util
 import json
 import subprocess
 import sys
 import urllib.request

 SIGNIN = "/home/super/Projects/NetVM/bin/muse-signin.py"
-CDP_PORTS = {"muse": "9410", "pip": "9420"}
+
+
+def _load_registry():
+    path = "/home/super/Projects/NetVM/bin/netvm-registry.py"
+    spec = importlib.util.spec_from_file_location("netvm_registry", path)
+    mod = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(mod)
+    return mod


 def cdp_ok(port):
@@ -51,9 +59,10 @@ def main():
        print("ERROR: unsupported service/id_type "
              "(muse+email only for now)", file=sys.stderr)
        return 1
-    port = CDP_PORTS.get(args.node)
+    port = _load_registry().port_for(args.node)
    if not port:
-        print("ERROR: unknown node", file=sys.stderr)
+        print("ERROR: unknown node '%s' (not in NODES.md registry)"
+              % args.node, file=sys.stderr)
        return 1

    if args.dry_run:
@@ -69,7 +78,7 @@ def main():
        print("ERROR: no identifier on stdin", file=sys.stderr)
        return 1

-    cmd = [sys.executable, SIGNIN, "--email", identifier]
+    cmd = [sys.executable, SIGNIN, "--node", args.node, "--email", identifier]
    if args.step == "submit":
        if not code:
            print("ERROR: no code on stdin", file=sys.stderr)