Esempio n. 1
0
// waitUntilFreshAndBlock waits until cache is at least as fresh as given <resourceVersion>.
// NOTE: This function acquired lock and doesn't release it.
// You HAVE TO explicitly call w.RUnlock() after this function.
func (w *watchCache) waitUntilFreshAndBlock(resourceVersion uint64, trace *util.Trace) error {
	startTime := w.clock.Now()
	go func() {
		// Wake us up when the time limit has expired.  The docs
		// promise that time.After (well, NewTimer, which it calls)
		// will wait *at least* the duration given. Since this go
		// routine starts sometime after we record the start time, and
		// it will wake up the loop below sometime after the broadcast,
		// we don't need to worry about waking it up before the time
		// has expired accidentally.
		<-w.clock.After(blockTimeout)
		w.cond.Broadcast()
	}()

	w.RLock()
	if trace != nil {
		trace.Step("watchCache locked acquired")
	}
	for w.resourceVersion < resourceVersion {
		if w.clock.Since(startTime) >= blockTimeout {
			// Timeout with retry after 1 second.
			return errors.NewTimeoutError(fmt.Sprintf("Too large resource version: %v, current: %v", resourceVersion, w.resourceVersion), 1)
		}
		w.cond.Wait()
	}
	if trace != nil {
		trace.Step("watchCache fresh enough")
	}
	return nil
}
Esempio n. 2
0
// finishRequest makes a given resultFunc asynchronous and handles errors returned by the response.
// Any api.Status object returned is considered an "error", which interrupts the normal response flow.
func finishRequest(timeout time.Duration, fn resultFunc) (result runtime.Object, err error) {
	// these channels need to be buffered to prevent the goroutine below from hanging indefinitely
	// when the select statement reads something other than the one the goroutine sends on.
	ch := make(chan runtime.Object, 1)
	errCh := make(chan error, 1)
	panicCh := make(chan interface{}, 1)
	go func() {
		// panics don't cross goroutine boundaries, so we have to handle ourselves
		defer utilruntime.HandleCrash(func(panicReason interface{}) {
			// Propagate to parent goroutine
			panicCh <- panicReason
		})

		if result, err := fn(); err != nil {
			errCh <- err
		} else {
			ch <- result
		}
	}()

	select {
	case result = <-ch:
		if status, ok := result.(*metav1.Status); ok {
			return nil, errors.FromObject(status)
		}
		return result, nil
	case err = <-errCh:
		return nil, err
	case p := <-panicCh:
		panic(p)
	case <-time.After(timeout):
		return nil, errors.NewTimeoutError("request did not complete within allowed duration", 0)
	}
}
Esempio n. 3
0
// Create attempts to create a new eviction.  That is, it tries to evict a pod.
func (r *EvictionREST) Create(ctx genericapirequest.Context, obj runtime.Object) (runtime.Object, error) {
	eviction := obj.(*policy.Eviction)

	obj, err := r.store.Get(ctx, eviction.Name, &metav1.GetOptions{})
	if err != nil {
		return nil, err
	}
	pod := obj.(*api.Pod)
	var rtStatus *metav1.Status
	var pdbName string
	err = retry.RetryOnConflict(EvictionsRetry, func() error {
		pdbs, err := r.getPodDisruptionBudgets(ctx, pod)
		if err != nil {
			return err
		}

		if len(pdbs) > 1 {
			rtStatus = &metav1.Status{
				Status:  metav1.StatusFailure,
				Message: "This pod has more than one PodDisruptionBudget, which the eviction subresource does not support.",
				Code:    500,
			}
			return nil
		} else if len(pdbs) == 1 {
			pdb := pdbs[0]
			pdbName = pdb.Name
			// Try to verify-and-decrement

			// If it was false already, or if it becomes false during the course of our retries,
			// raise an error marked as a 429.
			ok, err := r.checkAndDecrement(pod.Namespace, pod.Name, pdb)
			if err != nil {
				return err
			}

			if !ok {
				rtStatus = &metav1.Status{
					Status: metav1.StatusFailure,
					// TODO(mml): Include some more details about why the eviction is disallowed.
					// Ideally any such text is generated by the DisruptionController (offline).
					Message: "Cannot evict pod as it would violate the pod's disruption budget.",
					Code:    429,
					// TODO(mml): Add a Retry-After header.  Once there are time-based
					// budgets, we can sometimes compute a sensible suggested value.  But
					// even without that, we can give a suggestion (10 minutes?) that
					// prevents well-behaved clients from hammering us.
				}
			}
		}
		return nil
	})
	if err == wait.ErrWaitTimeout {
		err = errors.NewTimeoutError(fmt.Sprintf("couldn't update PodDisruptionBudget %q due to conflicts", pdbName), 10)
	}
	if err != nil {
		return nil, err
	}

	if rtStatus != nil {
		return rtStatus, nil
	}

	// At this point there was either no PDB or we succeded in decrementing

	// Try the delete
	_, err = r.store.Delete(ctx, eviction.Name, eviction.DeleteOptions)
	if err != nil {
		return nil, err
	}

	// Success!
	return &metav1.Status{Status: metav1.StatusSuccess}, nil
}