Make sure we close the response body before retrying

Not doing so can keep the HTTP client code waiting forever, keeping some goroutines running, and the socket open. Signed-off-by: Miloslav Trmač <[email protected]>
containers · Apr 5, 2024 · 77f9d7b · 77f9d7b
1 parent 489cf63
commit 77f9d7b
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/docker/docker_client.go b/docker/docker_client.go
@@ -578,6 +578,7 @@ func (c *dockerClient) makeRequestToResolvedURL(ctx context.Context, method stri
 		if attempts == 1 && stream == nil && auth != noAuth {
 			if retry, newScope := needsRetryWithUpdatedScope(res); retry {
 				logrus.Debug("Detected insufficient_scope error, will retry request with updated scope")
+				res.Body.Close()
 				// Note: This retry ignores extraScope. That’s, strictly speaking, incorrect, but we don’t currently
 				// expect the insufficient_scope errors to happen for those callers. If that changes, we can add support
 				// for more than one extra scope.