fix(worker): harden control-plane lifecycle
Все проверки выполнены успешно
CI / test (push) Successful in 2m32s
Docker / Build and publish worker image (push) Successful in 18m17s

- reconnect safely after token rotation and retry leased results
- reject malformed tasks and remove production cluster debug mutation
- validate environment files and require immutable container images

BREAKING CHANGE: Docker install, deploy, and Compose now require an
immutable repository@sha256 image reference.
Этот коммит содержится в:
Gleb Tv
2026-07-19 23:11:43 +03:00
родитель 6937674449
Коммит e987f24903
38 изменённых файлов: 2203 добавлений и 674 удалений

Просмотреть файл

@@ -74,63 +74,6 @@ func (s *Server) handleClusterStatus(w http.ResponseWriter, _ *http.Request) {
}
}
// handleClusterApplyTestConfig applies a hardcoded config.adopt log
// entry to the cluster. It exists so the e2e script and any operator
// debugging session can verify FSM replication without having to wire
// up the real signed-config-adoption producer (which lives in a later
// phase).
//
// DEBUG: this endpoint is a placeholder for the real producer. It must
// be replaced (or removed) before any production deployment.
//
// The handler is gated behind Config.DebugClusterApply (env
// WORKER_CLUSTER_DEBUG_APPLY=true). When the flag is false the
// handler returns 404 — the route is still registered so the auth
// + CSRF paths are exercised in tests, but no real FSM entry is ever
// appended from a production webapp.
//
// TODO(worker-cluster-real-producer): remove the apply-test-config
// endpoint entirely once the signed-config-adoption producer ships.
func (s *Server) handleClusterApplyTestConfig(w http.ResponseWriter, r *http.Request) {
writeNoStore(w)
if !s.cfg.DebugClusterApply {
http.NotFound(w, r)
return
}
if s.cluster == nil {
http.Error(w, "cluster not configured", http.StatusServiceUnavailable)
return
}
if !s.requireCSRF(sessionFromContextOrFail(w, r), r) {
http.Error(w, "csrf token required", http.StatusForbidden)
return
}
applied, err := s.cluster.ApplyTestConfig()
if err != nil {
s.deps.Logger.Printf("cluster apply test config: %v", err)
http.Error(w, err.Error(), http.StatusBadGateway)
return
}
w.Header().Set("Content-Type", "application/json; charset=utf-8")
w.WriteHeader(http.StatusOK)
if err := json.NewEncoder(w).Encode(map[string]uint64{"applied_index": applied}); err != nil {
s.deps.Logger.Printf("cluster apply encode: %v", err)
}
}
// sessionFromContextOrFail is a tiny adapter so requireCSRF can be
// called from this handler without leaking the middleware into the
// cluster package. If no session is attached (should not happen
// because requireSession already ran) we return a stub session with
// no CSRF token, which causes requireCSRF to refuse the request.
func sessionFromContextOrFail(_ http.ResponseWriter, r *http.Request) *Session {
sess, _ := sessionFromContext(r.Context())
if sess != nil {
return sess
}
return &Session{}
}
// ErrClusterNotConfigured is returned when a cluster-admin endpoint is
// hit on a server without a cluster attached.
var ErrClusterNotConfigured = errors.New("webapp: cluster not configured")