helm: retry on connection refused (#1245)

* bootstrapper: directly return kubewaiter error

* helm: retry on connection refused
This commit is contained in:
3u13r 2023-02-22 09:58:28 +01:00 committed by GitHub
parent d78d22f95a
commit 3339ae2399
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 9 additions and 7 deletions

View File

@ -15,6 +15,7 @@ import (
"net" "net"
"os/exec" "os/exec"
"strconv" "strconv"
"strings"
"time" "time"
"github.com/edgelesssys/constellation/v2/bootstrapper/internal/kubernetes/k8sapi" "github.com/edgelesssys/constellation/v2/bootstrapper/internal/kubernetes/k8sapi"
@ -188,6 +189,11 @@ func (h *Client) installCiliumGCP(ctx context.Context, kubectl k8sapi.Client, re
// The function will wait 30 seconds before retrying a failed installation attempt. // The function will wait 30 seconds before retrying a failed installation attempt.
// After 5 minutes the retrier will be canceld and the function returns with an error. // After 5 minutes the retrier will be canceld and the function returns with an error.
func (h *Client) install(ctx context.Context, chartRaw []byte, values map[string]any) error { func (h *Client) install(ctx context.Context, chartRaw []byte, values map[string]any) error {
retriable := func(err error) bool {
return errors.Is(err, wait.ErrWaitTimeout) ||
strings.Contains(err.Error(), "connection refused")
}
reader := bytes.NewReader(chartRaw) reader := bytes.NewReader(chartRaw)
chart, err := loader.LoadArchive(reader) chart, err := loader.LoadArchive(reader)
if err != nil { if err != nil {
@ -200,7 +206,7 @@ func (h *Client) install(ctx context.Context, chartRaw []byte, values map[string
values, values,
h.log, h.log,
} }
retrier := retry.NewIntervalRetrier(doer, 30*time.Second, isTimeoutErr) retrier := retry.NewIntervalRetrier(doer, 30*time.Second, retriable)
// Since we have no precise retry condition we want to stop retrying after 5 minutes. // Since we have no precise retry condition we want to stop retrying after 5 minutes.
// The helm library only reports a timeout error in the error cases we currently know. // The helm library only reports a timeout error in the error cases we currently know.
@ -229,8 +235,3 @@ func (i installDoer) Do(ctx context.Context) error {
return err return err
} }
// isTimeoutErr checks if the given error is a timeout error from k8s.io/apimachinery.
func isTimeoutErr(err error) bool {
return errors.Is(err, wait.ErrWaitTimeout)
}

View File

@ -9,6 +9,7 @@ package kubewaiter
import ( import (
"context" "context"
"fmt"
"time" "time"
corev1 "k8s.io/api/core/v1" corev1 "k8s.io/api/core/v1"
@ -33,7 +34,7 @@ func (w *CloudKubeAPIWaiter) Wait(ctx context.Context, kubernetesClient Kubernet
doer := &kubeDoer{kubeClient: kubernetesClient} doer := &kubeDoer{kubeClient: kubernetesClient}
retrier := retry.NewIntervalRetrier(doer, 5*time.Second, funcAlwaysRetriable) retrier := retry.NewIntervalRetrier(doer, 5*time.Second, funcAlwaysRetriable)
if err := retrier.Do(ctx); err != nil { if err := retrier.Do(ctx); err != nil {
return doer.Do(context.Background()) return fmt.Errorf("waiting for Kubernetes API: %w", err)
} }
return nil return nil
} }