fix: clear consumed ephemeral registration state (#1256)

Ephemeral registrations are single-use, but the daemon kept `.runner` after the task, so a restarted container failed `Declare` with `unregistered runner` forever.

An ephemeral runner now removes its registration file as soon as it fetches a task, so a container killed mid-job registers afresh on restart. It also removes the file when Gitea answers `unregistered runner` at startup or while polling. A proxy's 401 does not count. Persistent runners keep their file, because Gitea sends the same reply for a token mismatch.

Closes https://gitea.com/gitea/runner/issues/818

---------

Co-authored-by: silverwind <[email protected]>
Reviewed-on: https://gitea.com/gitea/runner/pulls/1256
Reviewed-by: silverwind <[email protected]>
This commit is contained in:
bircniandsilverwind committed 2026-10-01 09:23:45 +00:00
1 parent 0cedd98051
commit d93b7d278a
6 files changed
+148 -15

No files matched your search

+19 -1
View File
@@ -191,6 +191,9 @@ func runDaemon(ctx context.Context, daemArgs *daemonArgs, configFile *string) fu
return err
case err != nil:
log.WithError(err).Error("fail to invoke Declare")
if client.IsRegistrationRejected(err) {
removeEphemeralRegistration(reg, cfg.Runner.File)
}
return err
default:
log.Infof("runner: %s, with version: %s, with labels: %v, declare successfully",
@@ -215,7 +218,7 @@ func runDaemon(ctx context.Context, daemArgs *daemonArgs, configFile *string) fu
done := make(chan struct{})
go func() {
defer close(done)
poller.PollOnce()
poller.PollOnce(func() { removeEphemeralRegistration(reg, cfg.Runner.File) })
}()
// shutdown when we complete a job or cancel is requested
@@ -244,6 +247,10 @@ func runDaemon(ctx context.Context, daemArgs *daemonArgs, configFile *string) fu
log.Warnf("runner: %s cancelled in progress jobs during shutdown", resp.Msg.Runner.Name)
}
if poller.RegistrationRejected() {
removeEphemeralRegistration(reg, cfg.Runner.File)
}
if poller.Unregistered() {
return errors.New("runner is no longer registered with the server; please register it again")
}
@@ -252,6 +259,17 @@ func runDaemon(ctx context.Context, daemArgs *daemonArgs, configFile *string) fu
}
}
func removeEphemeralRegistration(reg *config.Registration, file string) {
if !reg.Ephemeral {
return
}
if err := os.Remove(file); err == nil {
log.Infof("removed the spent ephemeral registration %q", file)
} else if !os.IsNotExist(err) {
log.WithError(err).Warn("failed to remove the spent ephemeral registration")
}
}
type daemonArgs struct {
Once bool
Labels string
+87
View File
@@ -4,13 +4,24 @@
package cmd
import (
"context"
"errors"
"net/http"
"net/http/httptest"
"os"
"path/filepath"
"testing"
"time"
"gitea.com/gitea/runner/internal/pkg/config"
"gitea.com/gitea/runner/internal/pkg/report"
"connectrpc.com/connect"
runnerv1 "gitea.dev/actionslib/runner/v1"
"gitea.dev/actionslib/runner/v1/runnerv1connect"
log "github.com/sirupsen/logrus"
"github.com/stretchr/testify/require"
"google.golang.org/protobuf/types/known/structpb"
)
func TestResolveLabels(t *testing.T) {
@@ -70,3 +81,79 @@ func TestInitLoggingSetsLevelAndCaller(t *testing.T) {
// act plans a job on this logger, so a live task's secrets have to be masked out of it
require.IsType(t, report.MaskingFormatter(nil), log.StandardLogger().Formatter)
}
var errUnregisteredRunner = errors.New("rpc error: code = Unauthenticated desc = unregistered runner")
type daemonTestService struct {
runnerv1connect.UnimplementedRunnerServiceHandler
declareErr, fetchErr error
beforeTask bool
cancel context.CancelFunc
}
func (s *daemonTestService) Declare(context.Context, *connect.Request[runnerv1.DeclareRequest]) (*connect.Response[runnerv1.DeclareResponse], error) {
if s.declareErr != nil {
return nil, s.declareErr
}
return connect.NewResponse(&runnerv1.DeclareResponse{Runner: &runnerv1.Runner{Name: "test"}}), nil
}
func (s *daemonTestService) FetchTask(context.Context, *connect.Request[runnerv1.FetchTaskRequest]) (*connect.Response[runnerv1.FetchTaskResponse], error) {
if s.fetchErr != nil {
return nil, s.fetchErr
}
if s.beforeTask {
s.cancel()
return nil, context.Canceled
}
return connect.NewResponse(&runnerv1.FetchTaskResponse{Task: &runnerv1.Task{Id: 1, Context: &structpb.Struct{}, WorkflowPayload: []byte("invalid: [")}}), nil
}
func (*daemonTestService) UpdateTask(_ context.Context, req *connect.Request[runnerv1.UpdateTaskRequest]) (*connect.Response[runnerv1.UpdateTaskResponse], error) {
return connect.NewResponse(&runnerv1.UpdateTaskResponse{State: req.Msg.State}), nil
}
func (*daemonTestService) UpdateLog(_ context.Context, req *connect.Request[runnerv1.UpdateLogRequest]) (*connect.Response[runnerv1.UpdateLogResponse], error) {
return connect.NewResponse(&runnerv1.UpdateLogResponse{AckIndex: req.Msg.Index + int64(len(req.Msg.Rows))}), nil
}
func TestDaemonRemovesOnlyConsumedEphemeralRegistration(t *testing.T) {
for _, tc := range []struct {
name string
ephemeral, beforeTask bool
declareErr, fetchErr error
removed bool
}{
{"ephemeral failed task", true, false, nil, nil, true},
{"ephemeral interrupted before task", true, true, nil, nil, false},
{"persistent once", false, false, nil, nil, false},
{"ephemeral rejected on fetch", true, false, nil, errUnregisteredRunner, true},
{"ephemeral proxy 401 on fetch", true, false, nil, connect.NewError(connect.CodeUnauthenticated, errors.New("HTTP status 401 Unauthorized")), false},
{"ephemeral rejected on declare", true, false, errUnregisteredRunner, nil, true},
{"persistent rejected on declare", false, false, errUnregisteredRunner, nil, false},
} {
t.Run(tc.name, func(t *testing.T) {
ctx, cancel := context.WithTimeout(t.Context(), 5*time.Second)
defer cancel()
_, handler := runnerv1connect.NewRunnerServiceHandler(&daemonTestService{declareErr: tc.declareErr, fetchErr: tc.fetchErr, beforeTask: tc.beforeTask, cancel: cancel})
server := httptest.NewServer(http.StripPrefix("/api/actions", handler))
defer server.Close()
dir := t.TempDir()
regFile := filepath.Join(dir, "registration.json")
configFile := filepath.Join(dir, "config.yaml")
require.NoError(t, os.WriteFile(configFile, []byte("runner:\n file: "+regFile+"\n idle_cleanup_interval: 0s\ncache:\n enabled: false\n"), 0o600))
require.NoError(t, config.SaveRegistration(regFile, &config.Registration{Address: server.URL, UUID: "test", Name: "test", Token: "test", Labels: []string{"host:host"}, Ephemeral: tc.ephemeral}))
err := runDaemon(ctx, &daemonArgs{Once: true}, &configFile)(nil, nil)
if tc.declareErr == nil && tc.fetchErr == nil {
require.NoError(t, err)
} else {
require.Error(t, err)
}
if tc.removed {
require.NoFileExists(t, regFile)
} else {
require.FileExists(t, regFile)
}
})
}
}