Fix coordinator worker integration
This commit is contained in:
@@ -117,8 +117,10 @@ const DefaultMaxAttempts = 3
|
||||
func (t *Task) CanRetry() bool { return t.Attempt < t.MaxAttempts }
|
||||
|
||||
// IsLeaseHeldBy reports whether worker currently holds this task at attempt.
|
||||
func (t *Task) IsLeaseHeldBy(worker string, attempt int) bool {
|
||||
return t.LeaseOwner != nil && *t.LeaseOwner == worker && t.Attempt == attempt
|
||||
func (t *Task) IsLeaseHeldBy(worker string, attempt int, now time.Time) bool {
|
||||
return t.LeaseOwner != nil && t.LeaseExpiresAt != nil && now.Before(*t.LeaseExpiresAt) &&
|
||||
*t.LeaseOwner == worker && t.Attempt == attempt &&
|
||||
(t.Status == TaskLeased || t.Status == TaskRunning)
|
||||
}
|
||||
|
||||
// AsClaimed projects the task into the trimmed view handed to a worker:
|
||||
@@ -146,7 +148,7 @@ func (t *Task) AsClaimed() ClaimedTask {
|
||||
|
||||
// verifyLease is the guard every worker-driven transition shares: the caller
|
||||
// must own the lease and reference the attempt it was granted.
|
||||
func (t *Task) verifyLease(worker string, attempt int) error {
|
||||
func (t *Task) verifyLease(worker string, attempt int, now time.Time) error {
|
||||
// A task is worker-owned while leased or running: the first heartbeat moves
|
||||
// it from leased to running, but ownership rules are identical for both.
|
||||
if t.Status != TaskLeased && t.Status != TaskRunning {
|
||||
@@ -158,13 +160,16 @@ func (t *Task) verifyLease(worker string, attempt int) error {
|
||||
if t.Attempt != attempt {
|
||||
return ErrStaleAttempt
|
||||
}
|
||||
if t.LeaseExpiresAt == nil || !now.Before(*t.LeaseExpiresAt) {
|
||||
return ErrLeaseConflict
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// RenewLease extends the lease of the worker that holds it. The first heartbeat
|
||||
// also acknowledges start, moving the task from leased to running.
|
||||
func (t *Task) RenewLease(worker string, attempt int, until time.Time) error {
|
||||
if err := t.verifyLease(worker, attempt); err != nil {
|
||||
func (t *Task) RenewLease(worker string, attempt int, now, until time.Time) error {
|
||||
if err := t.verifyLease(worker, attempt, now); err != nil {
|
||||
return err
|
||||
}
|
||||
t.LeaseExpiresAt = &until
|
||||
@@ -195,7 +200,7 @@ func (t *Task) CompleteWith(resultArtifactID uuid.UUID, metrics map[string]any,
|
||||
return ErrResultConflict
|
||||
}
|
||||
|
||||
if err := t.verifyLease(worker, attempt); err != nil {
|
||||
if err := t.verifyLease(worker, attempt, now); err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
@@ -214,7 +219,7 @@ func (t *Task) CompleteWith(resultArtifactID uuid.UUID, metrics map[string]any,
|
||||
// Fail records a worker-reported failure. A retryable failure with attempts
|
||||
// left returns the task to the queue; otherwise it terminates as failed.
|
||||
func (t *Task) Fail(worker string, attempt int, code, message string, retryable bool, now time.Time) error {
|
||||
if err := t.verifyLease(worker, attempt); err != nil {
|
||||
if err := t.verifyLease(worker, attempt, now); err != nil {
|
||||
return err
|
||||
}
|
||||
t.ErrorCode = &code
|
||||
|
||||
@@ -172,14 +172,14 @@ func TestFirstHeartbeatMovesLeasedToRunning(t *testing.T) {
|
||||
task := leasedTask(1, 3)
|
||||
until := testLater.Add(time.Hour)
|
||||
|
||||
if err := task.RenewLease(testWorker, 1, until); err != nil {
|
||||
if err := task.RenewLease(testWorker, 1, testNow, until); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if task.Status != TaskRunning {
|
||||
t.Errorf("status = %q, want running after first heartbeat", task.Status)
|
||||
}
|
||||
// A second heartbeat keeps it running.
|
||||
if err := task.RenewLease(testWorker, 1, until); err != nil {
|
||||
if err := task.RenewLease(testWorker, 1, testNow, until); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if task.Status != TaskRunning {
|
||||
@@ -190,15 +190,15 @@ func TestFirstHeartbeatMovesLeasedToRunning(t *testing.T) {
|
||||
func TestRunningTaskCanBeCompletedAndExpired(t *testing.T) {
|
||||
// Complete works from running.
|
||||
task := leasedTask(1, 3)
|
||||
_ = task.RenewLease(testWorker, 1, testLater) // -> running
|
||||
_ = task.RenewLease(testWorker, 1, testNow, testLater) // -> running
|
||||
if err := task.CompleteWith(testResult, nil, testWorker, 1, testNow); err != nil {
|
||||
t.Errorf("complete from running: %v", err)
|
||||
}
|
||||
|
||||
// Expire reclaims a running task too.
|
||||
task2 := leasedTask(1, 3)
|
||||
_ = task2.RenewLease(testWorker, 1, testLater) // -> running
|
||||
task2.ExpireLease(testNow)
|
||||
_ = task2.RenewLease(testWorker, 1, testNow, testLater) // -> running
|
||||
task2.ExpireLease(testLater)
|
||||
if task2.Status != TaskPending {
|
||||
t.Errorf("status = %q, want pending after a running lease expires", task2.Status)
|
||||
}
|
||||
@@ -208,14 +208,29 @@ func TestRenewLeaseExtendsOnlyForHolder(t *testing.T) {
|
||||
task := leasedTask(1, 3)
|
||||
until := testLater.Add(time.Hour)
|
||||
|
||||
if err := task.RenewLease(testWorker, 1, until); err != nil {
|
||||
if err := task.RenewLease(testWorker, 1, testNow, until); err != nil {
|
||||
t.Fatalf("unexpected error: %v", err)
|
||||
}
|
||||
if !task.LeaseExpiresAt.Equal(until) {
|
||||
t.Error("lease must be extended")
|
||||
}
|
||||
|
||||
if err := task.RenewLease("worker-2", 1, until); !errors.Is(err, ErrLeaseConflict) {
|
||||
if err := task.RenewLease("worker-2", 1, testNow, until); !errors.Is(err, ErrLeaseConflict) {
|
||||
t.Errorf("err = %v, want ErrLeaseConflict", err)
|
||||
}
|
||||
}
|
||||
|
||||
func TestExpiredLeaseRejectsRenewalCompletionAndFailure(t *testing.T) {
|
||||
task := leasedTask(1, 3)
|
||||
expired := testLater.Add(time.Nanosecond)
|
||||
|
||||
if err := task.RenewLease(testWorker, 1, expired, expired.Add(time.Minute)); !errors.Is(err, ErrLeaseConflict) {
|
||||
t.Errorf("renew expired lease: err = %v, want ErrLeaseConflict", err)
|
||||
}
|
||||
if err := task.CompleteWith(testResult, nil, testWorker, 1, expired); !errors.Is(err, ErrLeaseConflict) {
|
||||
t.Errorf("complete expired lease: err = %v, want ErrLeaseConflict", err)
|
||||
}
|
||||
if err := task.Fail(testWorker, 1, "timeout", "expired", true, expired); !errors.Is(err, ErrLeaseConflict) {
|
||||
t.Errorf("fail expired lease: err = %v, want ErrLeaseConflict", err)
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user