feat(federation): admission control on worker registration (S10)

Register() previously trusted a self-declared id and self-chosen token
from any caller, and let a second caller silently hijack an existing
worker id by re-registering it with a different token. Adds an optional
pre-shared AdmitToken (ORCHESTRA_FEDERATION_ADMIT_TOKEN) and requires a
same-id re-registration to present the existing worker's own token.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01W1rkJ2hBMybnJctPbcy4tT
This commit is contained in:
kami
2026-07-27 23:34:02 +04:00
parent fab9225a78
commit 8b4955a687
5 changed files with 85 additions and 13 deletions
+24 -6
View File
@@ -19,11 +19,16 @@ type Worker struct {
}
type Registry struct {
mu sync.Mutex
workers map[string]Worker
TTL time.Duration
OnOffline func(Worker)
cursors map[string]uint64
mu sync.Mutex
// AdmitToken, if set, is a pre-shared secret every registration must
// present (S10: registration previously accepted a self-declared id and
// self-chosen token from any caller — admission-control-free). Leave
// empty only for a deliberately open deployment.
AdmitToken string
workers map[string]Worker
TTL time.Duration
OnOffline func(Worker)
cursors map[string]uint64
}
func (r *Registry) init() {
@@ -37,13 +42,26 @@ func (r *Registry) init() {
r.cursors = map[string]uint64{}
}
}
func (r *Registry) Register(w Worker) error {
// Register admits a worker. admitToken must match r.AdmitToken whenever one
// is configured. Re-registering an ID that's already claimed requires that
// worker's own current token, so a caller can't self-declare someone else's
// id and hijack an existing worker's identity/capacity.
func (r *Registry) Register(w Worker, admitToken string) error {
if w.ID == "" {
return errors.New("worker id required")
}
if w.Token == "" {
return errors.New("worker token required")
}
r.mu.Lock()
defer r.mu.Unlock()
r.init()
if r.AdmitToken != "" && admitToken != r.AdmitToken {
return ErrUnauthorized
}
if existing, ok := r.workers[w.ID]; ok && existing.Token != w.Token {
return ErrUnauthorized
}
w.LastSeen = time.Now().UTC()
w.Online = true
r.workers[w.ID] = w
+23 -2
View File
@@ -7,7 +7,7 @@ import (
func TestCursorIsMonotonicAndAuthenticationIsRequired(t *testing.T) {
r := &Registry{}
if err := r.Register(Worker{ID: "workpc", Token: "secret"}); err != nil {
if err := r.Register(Worker{ID: "workpc", Token: "secret"}, ""); err != nil {
t.Fatal(err)
}
if err := r.Authenticate("workpc", "wrong"); err != ErrUnauthorized {
@@ -27,10 +27,31 @@ func TestCursorIsMonotonicAndAuthenticationIsRequired(t *testing.T) {
}
}
func TestRegisterRequiresAdmitTokenAndOwnToken(t *testing.T) {
r := &Registry{AdmitToken: "admit-secret"}
if err := r.Register(Worker{ID: "workpc", Token: "secret"}, "wrong"); err != ErrUnauthorized {
t.Fatalf("wrong admit token: got %v", err)
}
if err := r.Register(Worker{ID: "workpc", Token: "secret"}, "admit-secret"); err != nil {
t.Fatal(err)
}
// Re-registering the same id with a different token is a hijack
// attempt (S10), not a legitimate re-registration, and must be refused
// even with a valid admit token.
if err := r.Register(Worker{ID: "workpc", Token: "different"}, "admit-secret"); err != ErrUnauthorized {
t.Fatalf("hijack with different token: got %v", err)
}
// The same worker re-registering with its own token (e.g. after a
// restart) must still succeed.
if err := r.Register(Worker{ID: "workpc", Token: "secret"}, "admit-secret"); err != nil {
t.Fatalf("legitimate re-registration: %v", err)
}
}
func TestOfflineHookRunsOnceOnTransition(t *testing.T) {
called := make(chan Worker, 1)
r := &Registry{TTL: time.Millisecond, OnOffline: func(w Worker) { called <- w }}
if err := r.Register(Worker{ID: "workpc"}); err != nil {
if err := r.Register(Worker{ID: "workpc", Token: "secret"}, ""); err != nil {
t.Fatal(err)
}
r.mu.Lock()
@@ -80,7 +80,7 @@ func TestCrossMachineLeaseAnchorAndQuotaArePerHost(t *testing.T) {
// worker (spec §2.1): it must be registered and reachable before the
// router would ever consider leasing to it.
workers := &federation.Registry{}
if err := workers.Register(federation.Worker{ID: "workpc", Address: "workpc.mesh", Token: "secret"}); err != nil {
if err := workers.Register(federation.Worker{ID: "workpc", Address: "workpc.mesh", Token: "secret"}, ""); err != nil {
t.Fatal(err)
}
if err := workers.Heartbeat("workpc"); err != nil {