fix: proxy ListObjects request to one of the server based on hash(bucket) (#9881)

cmd/bucket-listobjects-handlers.go

@@ -17,6 +17,8 @@
 package cmd
 
 import (
+	"context"
+	"io"
 	"net/http"
 	"strings"
 
@@ -157,6 +159,10 @@ func (api objectAPIHandlers) ListObjectsV2MHandler(w http.ResponseWriter, r *htt
 		return
 	}
 
+	if proxyListRequest(ctx, w, r, bucket) {
+		return
+	}
+
 	listObjectsV2 := objectAPI.ListObjectsV2
 
 	// Inititate a list objects operation based on the input params.
@@ -231,6 +237,10 @@ func (api objectAPIHandlers) ListObjectsV2Handler(w http.ResponseWriter, r *http
 		return
 	}
 
+	if proxyListRequest(ctx, w, r, bucket) {
+		return
+	}
+
 	listObjectsV2 := objectAPI.ListObjectsV2
 
 	// Inititate a list objects operation based on the input params.
@@ -262,6 +272,43 @@ func (api objectAPIHandlers) ListObjectsV2Handler(w http.ResponseWriter, r *http
 	writeSuccessResponseXML(w, encodeResponse(response))
 }
 
+func getListEndpoint(bucket string) ListEndpoint {
+	return globalListEndpoints[crcHashMod(bucket, len(globalListEndpoints))]
+}
+
+// Proxy the list request to the right server.
+func proxyListRequest(ctx context.Context, w http.ResponseWriter, r *http.Request, bucket string) (success bool) {
+	if len(globalListEndpoints) == 0 {
+		return false
+	}
+	ep := getListEndpoint(bucket)
+	if ep.isLocal {
+		return false
+	}
+	ctx = r.Context()
+	outreq := r.Clone(ctx)
+	outreq.URL.Scheme = "http"
+	outreq.URL.Host = ep.host
+	outreq.URL.Path = r.URL.Path
+	outreq.Header.Add("Host", r.Host)
+	if globalIsSSL {
+		outreq.URL.Scheme = "https"
+	}
+	outreq.Host = r.Host
+	res, err := ep.t.RoundTrip(outreq)
+	if err != nil {
+		return false
+	}
+	for k, vv := range res.Header {
+		for _, v := range vv {
+			w.Header().Set(k, v)
+		}
+	}
+	w.WriteHeader(res.StatusCode)
+	io.Copy(w, res.Body)
+	return true
+}
+
 // ListObjectsV1Handler - GET Bucket (List Objects) Version 1.
 // --------------------------
 // This implementation of the GET operation returns some or all (up to 10000)
@@ -300,6 +347,10 @@ func (api objectAPIHandlers) ListObjectsV1Handler(w http.ResponseWriter, r *http
 		return
 	}
 
+	if proxyListRequest(ctx, w, r, bucket) {
+		return
+	}
+
 	listObjects := objectAPI.ListObjects
 
 	// Inititate a list objects operation based on the input params.

cmd/endpoint.go

@@ -17,8 +17,10 @@
 package cmd
 
 import (
+	"crypto/tls"
 	"fmt"
 	"net"
+	"net/http"
 	"net/url"
 	"path"
 	"path/filepath"
@@ -31,6 +33,7 @@ import (
 	"github.com/minio/minio-go/v6/pkg/set"
 	"github.com/minio/minio/cmd/config"
 	"github.com/minio/minio/cmd/logger"
+	"github.com/minio/minio/cmd/rest"
 	"github.com/minio/minio/pkg/env"
 	"github.com/minio/minio/pkg/mountinfo"
 )
@@ -46,6 +49,14 @@ const (
 	URLEndpointType
 )
 
+// ListEndpoint - endpoint used for list redirects
+// See proxyListRequest() for details.
+type ListEndpoint struct {
+	host    string
+	t       *http.Transport
+	isLocal bool
+}
+
 // Endpoint - any type of endpoint.
 type Endpoint struct {
 	*url.URL
@@ -708,6 +719,50 @@ func GetRemotePeers(endpointZones EndpointZones) []string {
 	return peerSet.ToSlice()
 }
 
+// GetListEndpoints - get all endpoints that can be used to proxy list request.
+func GetListEndpoints(endpointZones EndpointZones) ([]ListEndpoint, error) {
+	var listeps []ListEndpoint
+
+	listepExists := func(host string) bool {
+		for _, listep := range listeps {
+			if listep.host == host {
+				return true
+			}
+		}
+		return false
+	}
+
+	for _, ep := range endpointZones {
+		for _, endpoint := range ep.Endpoints {
+			if endpoint.Type() != URLEndpointType {
+				continue
+			}
+
+			host := endpoint.Host
+			if listepExists(host) {
+				continue
+			}
+			hostName, _, err := net.SplitHostPort(host)
+			if err != nil {
+				return nil, err
+			}
+			var tlsConfig *tls.Config
+			if globalIsSSL {
+				tlsConfig = &tls.Config{
+					ServerName: hostName,
+					RootCAs:    globalRootCAs,
+				}
+			}
+			listeps = append(listeps, ListEndpoint{
+				host,
+				newCustomHTTPTransport(tlsConfig, rest.DefaultRESTTimeout)(),
+				endpoint.IsLocal,
+			})
+		}
+	}
+	return listeps, nil
+}
+
 func updateDomainIPs(endPoints set.StringSet) {
 	ipList := set.NewStringSet()
 	for e := range endPoints {

cmd/globals.go

@@ -279,6 +279,7 @@ var (
 	// If writes to FS backend should be O_SYNC.
 	globalFSOSync bool
 
+	globalListEndpoints []ListEndpoint
 	// Add new variable global values here.
 )
 

cmd/server-main.go

@@ -396,6 +396,9 @@ func serverMain(ctx *cli.Context) {
 	globalRootCAs, err = config.GetRootCAs(globalCertsCADir.Get())
 	logger.FatalIf(err, "Failed to read root CAs (%v)", err)
 
+	globalListEndpoints, err = GetListEndpoints(globalEndpoints)
+	logger.FatalIf(err, "Invalid command line arguments")
+
 	globalMinioEndpoint = func() string {
 		host := globalMinioHost
 		if host == "" {