Files
orchestra/internal/federation/federation.go
T
kami 8b4955a687 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
2026-07-27 23:34:02 +04:00

138 lines
3.2 KiB
Go

package federation
import (
"errors"
"sync"
"time"
)
var ErrUnknownWorker = errors.New("unknown worker")
var ErrUnauthorized = errors.New("worker authentication failed")
type Worker struct {
ID string `json:"id"`
Address string `json:"address"`
Capacity int `json:"capacity"`
LastSeen time.Time `json:"last_seen"`
Online bool `json:"online"`
Token string `json:"-"`
}
type Registry struct {
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() {
if r.TTL <= 0 {
r.TTL = 90 * time.Second
}
if r.workers == nil {
r.workers = map[string]Worker{}
}
if r.cursors == nil {
r.cursors = map[string]uint64{}
}
}
// 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
if _, ok := r.cursors[w.ID]; !ok {
r.cursors[w.ID] = 0
}
return nil
}
func (r *Registry) Authenticate(id, token string) error {
r.mu.Lock()
defer r.mu.Unlock()
r.init()
w, ok := r.workers[id]
if !ok {
return ErrUnknownWorker
}
if w.Token == "" || token == "" || w.Token != token {
return ErrUnauthorized
}
return nil
}
func (r *Registry) Cursor(id string) (uint64, error) {
r.mu.Lock()
defer r.mu.Unlock()
r.init()
if _, ok := r.workers[id]; !ok {
return 0, ErrUnknownWorker
}
return r.cursors[id], nil
}
func (r *Registry) Ack(id string, cursor uint64) error {
r.mu.Lock()
defer r.mu.Unlock()
r.init()
if _, ok := r.workers[id]; !ok {
return ErrUnknownWorker
}
if cursor < r.cursors[id] {
return errors.New("cursor moved backwards")
}
r.cursors[id] = cursor
return nil
}
func (r *Registry) Heartbeat(id string) error {
r.mu.Lock()
defer r.mu.Unlock()
r.init()
w, ok := r.workers[id]
if !ok {
return ErrUnknownWorker
}
w.LastSeen = time.Now().UTC()
w.Online = true
r.workers[id] = w
return nil
}
func (r *Registry) Snapshot() []Worker {
r.mu.Lock()
defer r.mu.Unlock()
r.init()
now := time.Now().UTC()
out := make([]Worker, 0, len(r.workers))
for id, w := range r.workers {
wasOnline := w.Online
w.Online = now.Sub(w.LastSeen) <= r.TTL
r.workers[id] = w
if wasOnline && !w.Online && r.OnOffline != nil {
go r.OnOffline(w)
}
out = append(out, w)
}
return out
}