vercel · gsoltis · Feb 18, 2022 · Feb 18, 2022
diff --git a/cli/internal/cache/cache.go b/cli/internal/cache/cache.go
@@ -3,11 +3,10 @@ package cache
 
 import (
 	"fmt"
-	"sync"
-  "golang.org/x/sync/errgroup"
+
 	"github.com/vercel/turborepo/cli/internal/config"
 	"github.com/vercel/turborepo/cli/internal/ui"
-
+	"golang.org/x/sync/errgroup"
 )
 
 // Cache is abstracted way to cache/fetch previously run tasks
@@ -65,7 +64,7 @@ func (mplex cacheMultiplexer) Put(target string, key string, duration int, files
 // but it's hard to fix that without breaking the cache abstraction.
 func (mplex cacheMultiplexer) storeUntil(target string, key string, duration int, outputGlobs []string, stopAt int) error {
 	// Attempt to store on all caches simultaneously.
-	g := new(errgroup.Group)
+	g := &errgroup.Group{}
 	for i, cache := range mplex.caches {
 		if i == stopAt {
 			break

diff --git a/cli/internal/cache/cache_http.go b/cli/internal/cache/cache_http.go
@@ -41,9 +41,6 @@ func (cache *httpCache) Put(target, hash string, duration int, files []string) e
 	// if cache.writable {
 	cache.requestLimiter.acquire()
 	defer cache.requestLimiter.release()
-	if cache.config.ApiClient.CurrentFailCount > cache.config.Cache.MaxRemoteFailCount {
-		return fmt.Errorf("skipping uploading artifacts to HTTP cache: too many failures")
-	}
 
 	r, w := io.Pipe()
 	go cache.write(w, hash, files)
@@ -106,9 +103,6 @@ func (cache *httpCache) storeFile(tw *tar.Writer, name string) error {
 func (cache *httpCache) Fetch(target, key string, _unusedOutputGlobs []string) (bool, []string, error) {
 	cache.requestLimiter.acquire()
 	defer cache.requestLimiter.release()
-	if cache.config.ApiClient.CurrentFailCount > cache.config.Cache.MaxRemoteFailCount {
-		return false, nil, fmt.Errorf("skipping downloading artifacts to HTTP cache: too many past failures")
-	}
 	m, files, err := cache.retrieve(key)
 	if err != nil {
 		return false, files, fmt.Errorf("failed to retrieve files from HTTP cache: %w", err)

diff --git a/cli/internal/client/client.go b/cli/internal/client/client.go
@@ -1,7 +1,10 @@
 package client
 
 import (
+	"context"
+	"crypto/x509"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"io"
 	"io/ioutil"
@@ -18,24 +21,30 @@ import (
 
 type ApiClient struct {
 	// The api's base URL
-	baseUrl          string
-	Token            string
-	turboVersion     string
-	CurrentFailCount uint64
+	baseUrl      string
+	Token        string
+	turboVersion string
+	// Number of failed requests before we stop trying to upload/download artifacts to the remote cache
+	maxRemoteFailCount uint64
+	// Must be used via atomic package
+	currentFailCount uint64
 	// An http client
 	HttpClient *retryablehttp.Client
 }
 
+// ErrTooManyFailures is returned from remote cache API methods after `maxRemoteFailCount` errors have occurred
+var ErrTooManyFailures = errors.New("skipping HTTP Request, too many failures have occurred")
+
 func (api *ApiClient) SetToken(token string) {
 	api.Token = token
 }
 
 // New creates a new ApiClient
-func NewClient(baseUrl string, logger hclog.Logger, turboVersion string) *ApiClient {
+func NewClient(baseUrl string, logger hclog.Logger, turboVersion string, maxRemoteFailCount uint64) *ApiClient {
 	client := &ApiClient{
-		baseUrl:          baseUrl,
-		turboVersion:     turboVersion,
-		CurrentFailCount: 0,
+		baseUrl:            baseUrl,
+		turboVersion:       turboVersion,
+		maxRemoteFailCount: maxRemoteFailCount,
 		HttpClient: &retryablehttp.Client{
 			HTTPClient: &http.Client{
 				Timeout: time.Duration(20 * time.Second),
@@ -47,27 +56,26 @@ func NewClient(baseUrl string, logger hclog.Logger, turboVersion string) *ApiCli
 			Logger:       logger,
 		},
 	}
+	client.HttpClient.CheckRetry = client.checkRetry
 	return client
 }
 
-func (client *ApiClient) retryCachePolicy(resp *http.Response, err error) (bool, error) {
+func (c *ApiClient) retryCachePolicy(resp *http.Response, err error) (bool, error) {
 	if err != nil {
-		if v, ok := err.(*url.Error); ok {
+		if errors.As(err, &x509.UnknownAuthorityError{}) {
 			// Don't retry if the error was due to TLS cert verification failure.
-			if _, ok := v.Err.(x509.UnknownAuthorityError); ok {
-				atomic.AddUint64(&client.CurrentFailCount, 1)
-				return false, v
-			}
+			atomic.AddUint64(&c.currentFailCount, 1)
+			return false, err
 		}
-		atomic.AddUint64(&client.CurrentFailCount, 1)
+		atomic.AddUint64(&c.currentFailCount, 1)
 		return true, nil
 	}
 
 	// 429 Too Many Requests is recoverable. Sometimes the server puts
 	// a Retry-After response header to indicate when the server is
 	// available to start processing request from client.
 	if resp.StatusCode == http.StatusTooManyRequests {
-		atomic.AddUint64(&client.CurrentFailCount, 1)
+		atomic.AddUint64(&c.currentFailCount, 1)
 		return true, nil
 	}
 
@@ -76,13 +84,41 @@ func (client *ApiClient) retryCachePolicy(resp *http.Response, err error) (bool,
 	// errors and may relate to outages on the server side. This will catch
 	// invalid response codes as well, like 0 and 999.
 	if resp.StatusCode == 0 || (resp.StatusCode >= 500 && resp.StatusCode != 501) {
-		atomic.AddUint64(&client.CurrentFailCount, 1)
+		atomic.AddUint64(&c.currentFailCount, 1)
 		return true, fmt.Errorf("unexpected HTTP status %s", resp.Status)
 	}
 
 	return false, fmt.Errorf("unexpected HTTP status %s", resp.Status)
 }
 
+func (c *ApiClient) checkRetry(ctx context.Context, resp *http.Response, err error) (bool, error) {
+	// do not retry on context.Canceled or context.DeadlineExceeded
+	if ctx.Err() != nil {
+		atomic.AddUint64(&c.currentFailCount, 1)
+		return false, ctx.Err()
+	}
+
+	// we're squashing the error from the request and substituting any error that might come
+	// from our retry policy.
+	shouldRetry, err := c.retryCachePolicy(resp, err)
+	if shouldRetry {
+		// Our policy says it's ok to retry, but we need to check the failure count
+		if retryErr := c.okToRequest(); retryErr != nil {
+			return false, retryErr
+		}
+	}
+	return shouldRetry, err
+}
+
+// okToRequest returns nil if it's ok to make a request, and returns the error to
+// return to the caller if a request is not allowed
+func (c *ApiClient) okToRequest() error {
+	if atomic.LoadUint64(&c.currentFailCount) < c.maxRemoteFailCount {
+		return nil
+	}
+	return ErrTooManyFailures
+}
+
 func (c *ApiClient) makeUrl(endpoint string) string {
 	return fmt.Sprintf("%v%v", c.baseUrl, endpoint)
 }
@@ -92,6 +128,9 @@ func (c *ApiClient) UserAgent() string {
 }
 
 func (c *ApiClient) PutArtifact(hash string, teamId string, slug string, duration int, rawBody interface{}) error {
+	if err := c.okToRequest(); err != nil {
+		return err
+	}
 	params := url.Values{}
 	if teamId != "" && strings.HasPrefix(teamId, "team_") {
 		params.Add("teamId", teamId)
@@ -113,18 +152,6 @@ func (c *ApiClient) PutArtifact(hash string, teamId string, slug string, duratio
 		return fmt.Errorf("[WARNING] Invalid cache URL: %w", err)
 	}
 
-	c.HttpClient.CheckRetry = func(ctx context.Context, resp *http.Response, err error) (bool, error) {
-		// do not retry on context.Canceled or context.DeadlineExceeded
-		if ctx.Err() != nil {
-			c.CurrentFailCount = atomic.AddUint64(&c.CurrentFailCount, 1)
-			return false, ctx.Err()
-		}
-
-		// don't propagate other errors
-		shouldRetry, err := c.retryCachePolicy(resp, err)
-		return shouldRetry, err
-	}
-
 	if resp, err := c.HttpClient.Do(req); err != nil {
 		return fmt.Errorf("failed to store files in HTTP cache: %w", err)
 	} else {
@@ -134,6 +161,9 @@ func (c *ApiClient) PutArtifact(hash string, teamId string, slug string, duratio
 }
 
 func (c *ApiClient) FetchArtifact(hash string, teamId string, slug string, rawBody interface{}) (*http.Response, error) {
+	if err := c.okToRequest(); err != nil {
+		return nil, err
+	}
 	params := url.Values{}
 	if teamId != "" && strings.HasPrefix(teamId, "team_") {
 		params.Add("teamId", teamId)
@@ -153,18 +183,6 @@ func (c *ApiClient) FetchArtifact(hash string, teamId string, slug string, rawBo
 		return nil, fmt.Errorf("invalid cache URL: %w", err)
 	}
 
-	c.HttpClient.CheckRetry = func(ctx context.Context, resp *http.Response, err error) (bool, error) {
-		// do not retry on context.Canceled or context.DeadlineExceeded
-		if ctx.Err() != nil {
-			c.CurrentFailCount = atomic.AddUint64(&c.CurrentFailCount, 1)
-			return false, ctx.Err()
-		}
-
-		// don't propagate other errors
-		shouldRetry, err := c.retryCachePolicy(resp, err)
-		return shouldRetry, err
-	}
-
 	return c.HttpClient.Do(req)
 }
 

diff --git a/cli/internal/config/config.go b/cli/internal/config/config.go
@@ -9,6 +9,7 @@ import (
 	"path/filepath"
 	"runtime"
 	"strings"
+
 	"github.com/vercel/turborepo/cli/internal/client"
 
 	hclog "github.com/hashicorp/go-hclog"
@@ -49,8 +50,6 @@ type Config struct {
 
 // CacheConfig
 type CacheConfig struct {
-	// Number of failed requests before we stop trying to upload/download artifacts to the remote cache
-	MaxRemoteFailCount uint64
 	// Number of async workers
 	Workers int
 	// Cache directory
@@ -162,7 +161,8 @@ func ParseAndValidate(args []string, ui cli.Ui, turboVersion string) (c *Config,
 		Output: output,
 	})
 
-	apiClient := client.NewClient(partialConfig.ApiUrl, logger, turboVersion)
+	maxRemoteFailCount := 3
+	apiClient := client.NewClient(partialConfig.ApiUrl, logger, turboVersion, uint64(maxRemoteFailCount))
 
 	c = &Config{
 		Logger:       logger,
@@ -174,9 +174,8 @@ func ParseAndValidate(args []string, ui cli.Ui, turboVersion string) (c *Config,
 		ApiClient:    apiClient,
 		TurboVersion: turboVersion,
 		Cache: &CacheConfig{
-			MaxRemoteFailCount: 3,
-			Workers:            runtime.NumCPU() + 2,
-			Dir:                filepath.Join("node_modules", ".cache", "turbo"),
+			Workers: runtime.NumCPU() + 2,
+			Dir:     filepath.Join("node_modules", ".cache", "turbo"),
 		},
 	}