seaweedfs/weed/server/common.go

package weed_server

import (
	"bufio"
	"bytes"
	"context"
	"encoding/json"
	"errors"
	"fmt"
	"io"
	"io/fs"
	"mime"
	"mime/multipart"
	"net/http"
	"net/url"
	"path"
	"path/filepath"
	"strconv"
	"strings"
	"sync"
	"time"

	"github.com/seaweedfs/seaweedfs/weed/s3api/s3_constants"
	"github.com/seaweedfs/seaweedfs/weed/util/request_id"
	"google.golang.org/grpc/metadata"

	"github.com/seaweedfs/seaweedfs/weed/filer"

	"google.golang.org/grpc"

	"github.com/gorilla/mux"
	"github.com/seaweedfs/seaweedfs/weed/glog"
	"github.com/seaweedfs/seaweedfs/weed/operation"
	"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
	"github.com/seaweedfs/seaweedfs/weed/stats"
	"github.com/seaweedfs/seaweedfs/weed/storage/needle"
)

var serverStats *stats.ServerStats
var startTime = time.Now()
var writePool = sync.Pool{New: func() interface{} {
	return bufio.NewWriterSize(nil, 128*1024)
},
}

// ErrCacheNotReady signals that a remote-only object's local cache is still
// filling. Callers should map it to 503 + Retry-After so SDKs back off and retry.
var ErrCacheNotReady = errors.New("remote object not cached yet")

// writePrepareWriteFnErr writes an HTTP response for an error from
// prepareWriteFn, before any 2xx headers have been written. Client cancels are
// silent; filer_pb.ErrNotFound becomes 404; ErrCacheNotReady becomes 503 +
// Retry-After; everything else is 500. Strips headers that described the
// success body (Content-Length / Content-Range / Content-Disposition / ETag /
// Last-Modified) so they don't get attached to the error response.
func writePrepareWriteFnErr(w http.ResponseWriter, err error) {
	for _, h := range []string{"Content-Length", "Content-Range", "Content-Disposition", "ETag", "Last-Modified"} {
		w.Header().Del(h)
	}
	switch {
	case errors.Is(err, context.Canceled):
		glog.V(3).Infof("ProcessRangeRequest: client disconnected: %v", err)
	case errors.Is(err, filer_pb.ErrNotFound):
		http.Error(w, err.Error(), http.StatusNotFound)
	case errors.Is(err, ErrCacheNotReady):
		glog.V(1).Infof("ProcessRangeRequest: cache not ready, returning 503: %v", err)
		w.Header().Set("Retry-After", "5")
		http.Error(w, err.Error(), http.StatusServiceUnavailable)
	default:
		glog.Errorf("ProcessRangeRequest: %v", err)
		http.Error(w, err.Error(), http.StatusInternalServerError)
	}
}

func init() {
	serverStats = stats.NewServerStats()
	go serverStats.Start()
}

// bodyAllowedForStatus is a copy of http.bodyAllowedForStatus non-exported function.
func bodyAllowedForStatus(status int) bool {
	switch {
	case status >= 100 && status <= 199:
		return false
	case status == http.StatusNoContent:
		return false
	case status == http.StatusNotModified:
		return false
	}
	return true
}

func writeJson(w http.ResponseWriter, r *http.Request, httpStatus int, obj interface{}) (err error) {
	if !bodyAllowedForStatus(httpStatus) {
		return
	}

	var bytes []byte
	if obj != nil {
		if r.URL.Query().Get("pretty") != "" {
			bytes, err = json.MarshalIndent(obj, "", "  ")
		} else {
			bytes, err = json.Marshal(obj)
		}
	}
	if err != nil {
		return
	}

	if httpStatus >= 400 {
		glog.V(0).Infof("response method:%s URL:%s with httpStatus:%d and JSON:%s",
			r.Method, redactedURL(r), httpStatus, string(bytes))
	}

	w.Header().Set("Content-Type", "application/json")
	w.Header().Set("X-Content-Type-Options", "nosniff")
	w.WriteHeader(httpStatus)
	_, err = w.Write(bytes)
	return
}

func redactedURL(r *http.Request) string {
	if _, ok := r.URL.Query()["jwt"]; !ok {
		return r.URL.String()
	}
	u := *r.URL
	q := u.Query()
	q.Set("jwt", "[redacted]")
	u.RawQuery = q.Encode()
	return u.String()
}

// wrapper for writeJson - just logs errors
func writeJsonQuiet(w http.ResponseWriter, r *http.Request, httpStatus int, obj interface{}) {
	if err := writeJson(w, r, httpStatus, obj); err != nil {
		glog.V(0).Infof("error writing JSON status %s %d: %v", redactedURL(r), httpStatus, err)
		glog.V(1).Infof("JSON content: %+v", obj)
	}
}
func writeJsonError(w http.ResponseWriter, r *http.Request, httpStatus int, err error) {
	m := make(map[string]interface{})
	m["error"] = err.Error()
	glog.V(1).Infof("error JSON response status %d: %s", httpStatus, m["error"])
	writeJsonQuiet(w, r, httpStatus, m)
}

func debug(params ...interface{}) {
	glog.V(4).Infoln(params...)
}

func submitForClientHandler(w http.ResponseWriter, r *http.Request, masterFn operation.GetMasterFn, grpcDialOption grpc.DialOption, fileSizeLimitBytes int64) {
	ctx := r.Context()
	m := make(map[string]interface{})
	if r.Method != http.MethodPost {
		writeJsonError(w, r, http.StatusMethodNotAllowed, errors.New("Only submit via POST!"))
		return
	}

	debug("parsing upload file...")
	bytesBuffer := bufPool.Get().(*bytes.Buffer)
	defer bufPool.Put(bytesBuffer)
	pu, pe := needle.ParseUpload(r, fileSizeLimitBytes, bytesBuffer)
	if pe != nil {
		writeJsonError(w, r, http.StatusBadRequest, pe)
		return
	}

	debug("assigning file id for", pu.FileName)
	r.ParseForm()
	count := uint64(1)
	if r.FormValue("count") != "" {
		count, pe = strconv.ParseUint(r.FormValue("count"), 10, 32)
		if pe != nil {
			writeJsonError(w, r, http.StatusBadRequest, pe)
			return
		}
	}
	ar := &operation.VolumeAssignRequest{
		Count:            count,
		DataCenter:       r.FormValue("dataCenter"),
		Rack:             r.FormValue("rack"),
		Replication:      r.FormValue("replication"),
		Collection:       r.FormValue("collection"),
		Ttl:              r.FormValue("ttl"),
		DiskType:         r.FormValue("disk"),
		ExpectedDataSize: uint64(max(int64(0), int64(pu.OriginalDataSize))),
	}
	assignResult, ae := operation.Assign(ctx, masterFn, grpcDialOption, ar)
	if ae != nil {
		writeJsonError(w, r, http.StatusInternalServerError, ae)
		return
	}

	url := "http://" + assignResult.Url + "/" + assignResult.Fid
	if pu.ModifiedTime != 0 {
		url = url + "?ts=" + strconv.FormatUint(pu.ModifiedTime, 10)
	}

	debug("upload file to store", url)
	uploadOption := &operation.UploadOption{
		UploadUrl:         url,
		Filename:          pu.FileName,
		Cipher:            false,
		IsInputCompressed: pu.IsGzipped,
		MimeType:          pu.MimeType,
		PairMap:           pu.PairMap,
		Jwt:               assignResult.Auth,
	}
	uploader, err := operation.NewUploader()
	if err != nil {
		writeJsonError(w, r, http.StatusInternalServerError, err)
		return
	}
	uploadResult, err := uploader.UploadData(ctx, pu.Data, uploadOption)
	if err != nil {
		writeJsonError(w, r, http.StatusInternalServerError, err)
		return
	}

	m["fileName"] = pu.FileName
	m["fid"] = assignResult.Fid
	m["fileUrl"] = assignResult.PublicUrl + "/" + assignResult.Fid
	m["size"] = pu.OriginalDataSize
	m["eTag"] = uploadResult.ETag
	writeJsonQuiet(w, r, http.StatusCreated, m)
	return
}

func parseURLPath(path string) (vid, fid, filename, ext string, isVolumeIdOnly bool) {
	switch strings.Count(path, "/") {
	case 3:
		parts := strings.Split(path, "/")
		vid, fid, filename = parts[1], parts[2], parts[3]
		ext = filepath.Ext(filename)
	case 2:
		parts := strings.Split(path, "/")
		vid, fid = parts[1], parts[2]
		dotIndex := strings.LastIndex(fid, ".")
		if dotIndex > 0 {
			ext = fid[dotIndex:]
			fid = fid[0:dotIndex]
		}
	default:
		sepIndex := strings.LastIndex(path, "/")
		commaIndex := strings.LastIndex(path[sepIndex:], ",")
		if commaIndex <= 0 {
			vid, isVolumeIdOnly = path[sepIndex+1:], true
			return
		}
		dotIndex := strings.LastIndex(path[sepIndex:], ".")
		vid = path[sepIndex+1 : commaIndex]
		fid = path[commaIndex+1:]
		ext = ""
		if dotIndex > 0 {
			fid = path[commaIndex+1 : dotIndex]
			ext = path[dotIndex:]
		}
	}
	return
}

var StaticFS fs.FS

func handleStaticResources(defaultMux *http.ServeMux) {
	defaultMux.Handle("/favicon.ico", http.FileServer(http.FS(StaticFS)))
	defaultMux.Handle("/seaweedfsstatic/", http.StripPrefix("/seaweedfsstatic", http.FileServer(http.FS(StaticFS))))
}

func handleStaticResources2(r *mux.Router) {
	r.Handle("/favicon.ico", http.FileServer(http.FS(StaticFS)))
	r.PathPrefix("/seaweedfsstatic/").Handler(http.StripPrefix("/seaweedfsstatic", http.FileServer(http.FS(StaticFS))))
}

func AdjustPassthroughHeaders(w http.ResponseWriter, r *http.Request, filename string) {
	// Apply S3 passthrough headers from query parameters
	// AWS S3 supports overriding response headers via query parameters like:
	// ?response-cache-control=no-cache&response-content-type=application/json
	for queryParam, headerValue := range r.URL.Query() {
		if normalizedHeader, ok := s3_constants.PassThroughHeaders[strings.ToLower(queryParam)]; ok && len(headerValue) > 0 {
			w.Header().Set(normalizedHeader, headerValue[0])
		}
	}
	adjustHeaderContentDisposition(w, r, filename)
}
func adjustHeaderContentDisposition(w http.ResponseWriter, r *http.Request, filename string) {
	if contentDisposition := w.Header().Get("Content-Disposition"); contentDisposition != "" {
		return
	}
	if filename != "" {
		dispositionType := "inline"
		if r.FormValue("dl") != "" {
			if dl, _ := strconv.ParseBool(r.FormValue("dl")); dl {
				dispositionType = "attachment"
			}
		}
		// Use mime.FormatMediaType for RFC 6266 compliant Content-Disposition,
		// properly handling non-ASCII characters and special characters
		w.Header().Set("Content-Disposition", mime.FormatMediaType(dispositionType, map[string]string{"filename": filename}))
	}
}

// responseBodyTracker records how many bytes have reached the ResponseWriter.
// Bytes still sitting in the pooled bufio.Writer do not count: the status line
// is not committed until the buffer flushes.
type responseBodyTracker struct {
	http.ResponseWriter
	written int64
}

func (t *responseBodyTracker) Write(p []byte) (int, error) {
	n, err := t.ResponseWriter.Write(p)
	t.written += int64(n)
	return n, err
}

func ProcessRangeRequest(r *http.Request, w http.ResponseWriter, totalSize int64, mimeType string, prepareWriteFn func(offset int64, size int64) (filer.DoStreamContent, error)) error {
	rangeReq := r.Header.Get("Range")
	sink := &responseBodyTracker{ResponseWriter: w}
	bufferedWriter := writePool.Get().(*bufio.Writer)
	bufferedWriter.Reset(sink)
	discardBuffered := false
	defer func() {
		if discardBuffered {
			bufferedWriter.Reset(io.Discard)
		} else {
			bufferedWriter.Flush()
		}
		writePool.Put(bufferedWriter)
	}()

	if rangeReq == "" {
		w.Header().Set("Content-Length", strconv.FormatInt(totalSize, 10))
		writeFn, err := prepareWriteFn(0, totalSize)
		if err != nil {
			writePrepareWriteFnErr(w, err)
			return fmt.Errorf("ProcessRangeRequest: %w", err)
		}
		if err = writeFn(bufferedWriter); err != nil {
			glog.Errorf("ProcessRangeRequest: %v", err)
			// Drop the unflushed tail: flushing it would finish a 200 whose
			// Content-Length is already set, delivering corrupt bytes as a
			// complete body.
			discardBuffered = true
			if sink.written > 0 {
				panic(http.ErrAbortHandler)
			}
			w.Header().Del("Content-Length")
			http.Error(w, err.Error(), http.StatusInternalServerError)
			return fmt.Errorf("ProcessRangeRequest: %w", err)
		}
		return nil
	}

	//the rest is dealing with partial content request
	//mostly copy from src/pkg/net/http/fs.go
	ranges, err := parseRange(rangeReq, totalSize)
	if err != nil {
		glog.Errorf("ProcessRangeRequest headers: %+v err: %v", w.Header(), err)
		if err == errNoOverlap {
			w.Header().Set("Content-Range", fmt.Sprintf("bytes */%d", totalSize))
		}
		http.Error(w, err.Error(), http.StatusRequestedRangeNotSatisfiable)
		return fmt.Errorf("ProcessRangeRequest header: %w", err)
	}
	if sumRangesSize(ranges) > totalSize {
		// The total number of bytes in all the ranges
		// is larger than the size of the file by
		// itself, so this is probably an attack, or a
		// dumb client.  Ignore the range request.
		return nil
	}
	if len(ranges) == 0 {
		return nil
	}
	if len(ranges) == 1 {
		// RFC 2616, Section 14.16:
		// "When an HTTP message includes the content of a single
		// range (for example, a response to a request for a
		// single range, or to a request for a set of ranges
		// that overlap without any holes), this content is
		// transmitted with a Content-Range header, and a
		// Content-Length header showing the number of bytes
		// actually transferred.
		// ...
		// A response to a request for a single range MUST NOT
		// be sent using the multipart/byteranges media type."
		ra := ranges[0]
		w.Header().Set("Content-Length", strconv.FormatInt(ra.length, 10))
		w.Header().Set("Content-Range", ra.contentRange(totalSize))

		writeFn, err := prepareWriteFn(ra.start, ra.length)
		if err != nil {
			writePrepareWriteFnErr(w, err)
			return fmt.Errorf("ProcessRangeRequest: %w", err)
		}
		w.WriteHeader(http.StatusPartialContent)
		err = writeFn(bufferedWriter)
		if err != nil {
			glog.Errorf("ProcessRangeRequest range[0]: %+v err: %v", w.Header(), err)
			// WriteHeader was already called: drop the unflushed tail and
			// abort so the client does not read corrupt bytes as a full body.
			discardBuffered = true
			panic(http.ErrAbortHandler)
		}
		return nil
	}

	// process multiple ranges
	writeFnByRange := make(map[int](func(writer io.Writer) error))

	for i, ra := range ranges {
		if ra.start > totalSize {
			http.Error(w, "Out of Range", http.StatusRequestedRangeNotSatisfiable)
			return fmt.Errorf("out of range: %w", err)
		}
		writeFn, err := prepareWriteFn(ra.start, ra.length)
		if err != nil {
			writePrepareWriteFnErr(w, err)
			return fmt.Errorf("ProcessRangeRequest range[%d]: %w", i, err)
		}
		writeFnByRange[i] = writeFn
	}
	sendSize := rangesMIMESize(ranges, mimeType, totalSize)
	pr, pw := io.Pipe()
	mw := multipart.NewWriter(pw)
	w.Header().Set("Content-Type", "multipart/byteranges; boundary="+mw.Boundary())
	sendContent := pr
	defer pr.Close() // cause writing goroutine to fail and exit if CopyN doesn't finish.
	go func() {
		for i, ra := range ranges {
			part, e := mw.CreatePart(ra.mimeHeader(mimeType, totalSize))
			if e != nil {
				pw.CloseWithError(e)
				return
			}
			writeFn := writeFnByRange[i]
			if writeFn == nil {
				pw.CloseWithError(e)
				return
			}
			if e = writeFn(part); e != nil {
				pw.CloseWithError(e)
				return
			}
		}
		mw.Close()
		pw.Close()
	}()
	if w.Header().Get("Content-Encoding") == "" {
		w.Header().Set("Content-Length", strconv.FormatInt(sendSize, 10))
	}
	w.WriteHeader(http.StatusPartialContent)
	if _, err := io.CopyN(bufferedWriter, sendContent, sendSize); err != nil {
		glog.Errorf("ProcessRangeRequest err: %v", err)
		// WriteHeader was already called: drop the unflushed tail and abort.
		discardBuffered = true
		panic(http.ErrAbortHandler)
	}
	return nil
}

// CleanPathHandler serves a request whose path is not canonical ("//", "." or
// ".." segments) at the cleaned path instead of letting http.ServeMux redirect
// to it. ServeMux builds that redirect from the already percent-encoded path, so
// the Location header is encoded twice (golang/go#79897): a client following it
// re-sends "/负极全景" as "/%25E8%25B4%259F...", and the filer then stores a
// directory literally named "%E8%B4%9F...". Cleaning here mirrors the path
// ServeMux would have redirected to, so the decoded name reaches the handler.
func CleanPathHandler(h http.Handler) http.Handler {
	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		escaped := r.URL.EscapedPath()
		if cleaned := cleanPath(escaped); cleaned != escaped {
			if p, err := url.PathUnescape(cleaned); err == nil {
				r2 := new(http.Request)
				*r2 = *r
				r2.URL = new(url.URL)
				*r2.URL = *r.URL
				r2.URL.Path, r2.URL.RawPath = p, cleaned
				// PostHandler picks storage rules and the bucket from RequestURI, so
				// keep it in step with the path the entry is written to.
				r2.RequestURI = r2.URL.RequestURI()
				r = r2
			}
		}
		h.ServeHTTP(w, r)
	})
}

// cleanPath is the canonical form http.ServeMux redirects to: path.Clean plus
// the trailing slash, which the filer relies on to tell a directory from a file.
func cleanPath(p string) string {
	if p == "" {
		return "/"
	}
	if p[0] != '/' {
		p = "/" + p
	}
	np := path.Clean(p)
	if p[len(p)-1] == '/' && np != "/" {
		np += "/"
	}
	return np
}

func requestIDMiddleware(h http.HandlerFunc) http.HandlerFunc {
	return func(w http.ResponseWriter, r *http.Request) {
		request_id.Middleware(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
			ctx := metadata.NewOutgoingContext(r.Context(),
				metadata.New(map[string]string{
					request_id.AmzRequestIDHeader: request_id.Get(r.Context()),
				}))
			h(w, r.WithContext(ctx))
		})).ServeHTTP(w, r)
	}
}