Give the ping context a timeout.

mattmoor · mattmoor · commit ff62435e5e6b · 2021-11-03T17:38:24.000-07:00
Debugging why requests to an insecure registry seem to hang, I noticed that there are 3+ requests to the https endpoint before it falls back to http (at that point my logs truncate because the test was over).  The time between the request being issued and completing in the roundtripper was ~30s each.

This change gives each ping a generous 5s timeout to complete each request.
diff --git a/pkg/v1/remote/transport/ping.go b/pkg/v1/remote/transport/ping.go
@@ -22,6 +22,7 @@ import (
 	"io/ioutil"
 	"net/http"
 	"strings"
+	"time"
 
 	authchallenge "github.com/docker/distribution/registry/client/auth/challenge"
 	"github.com/google/go-containerregistry/pkg/name"
@@ -86,7 +87,12 @@ func ping(ctx context.Context, reg name.Registry, t http.RoundTripper) (*pingRes
 		if err != nil {
 			return nil, err
 		}
-		resp, err := client.Do(req.WithContext(ctx))
+		// The ping handler should be extremely fast, but for registries that serve
+		// over http, it could take a while to fallback from https (esp. if the
+		// transport has retries).  So give each ping attempt a (generous) 5s timeout.
+		pctx, cancel := context.WithTimeout(ctx, 5*time.Second)
+		defer cancel()
+		resp, err := client.Do(req.WithContext(pctx))
 		if err != nil {
 			errs = append(errs, err.Error())
 			// Potentially retry with http.
