nzinkevich/frostfs-http-gw
1
0
Fork 0
forked from TrueCloudLab/frostfs-http-gw
Code Pull requests Activity

[#170] Support .tar/.tgz unpacking during request

Browse source 
During upload if "X-Explode-Archive" is set, gate tries to read archive and creates an object for each file. Each object acquires a FilePath attribute which is calculated relative to the archive root

Signed-off-by: Nikita Zinkevich <n.zinkevich@yadro.com>
This commit is contained in:
Nikita Zinkevich 2024-12-06 15:01:16 +03:00
parent b17a1a8dd5
commit 72530c57ca
Signed by: nzinkevich
GPG key ID: 748EA1D0B2E6420A
5 changed files with 145 additions and 54 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
internal/handler/filter.go
View file

@ -39,7 +39,7 @@ func filterHeaders(l *zap.Logger, header *fasthttp.RequestHeader) (map[string]st
// check if key gets duplicated
// return error containing full key name (with prefix)
if _, ok := result[string(clearKey)]; ok {
err = fmt.Errorf("key duplication error: %s", string(key))
err = fmt.Errorf("header key duplication error: %s", string(key))
return
}

4
internal/handler/multipart.go
View file

@ -42,7 +42,9 @@ func fetchMultipartFile(l *zap.Logger, r io.Reader, boundary string) (MultipartF
// ignore multipart/form-data values
if filename == "" {
l.Debug(logs.IgnorePartEmptyFilename, zap.String("form", name))
if err = part.Close(); err != nil {
l.Warn(logs.FailedToCloseReader, zap.Error(err))
}
continue
}

184
internal/handler/upload.go
View file

@ -1,13 +1,19 @@
package handler
import (
"archive/tar"
"bytes"
"compress/gzip"
"context"
"encoding/json"
"errors"
"io"
"net/http"
"path/filepath"
"strconv"
"time"
"git.frostfs.info/TrueCloudLab/frostfs-http-gw/internal/data"
"git.frostfs.info/TrueCloudLab/frostfs-http-gw/internal/logs"
"git.frostfs.info/TrueCloudLab/frostfs-http-gw/response"
"git.frostfs.info/TrueCloudLab/frostfs-http-gw/tokens"
@ -20,10 +26,37 @@ import (
)
const (
jsonHeader = "application/json; charset=UTF-8"
drainBufSize = 4096
jsonHeader = "application/json; charset=UTF-8"
drainBufSize = 4096
explodeArchiveHeader = "X-Explode-Archive"
)
type accumulatedReader struct {
reader io.Reader
buffer *bytes.Buffer
}
func newAccumulatedReader(rc io.Reader) *accumulatedReader {
return &accumulatedReader{
reader: rc,
buffer: &bytes.Buffer{},
}
}
// Read reads data from the underlying io.Reader and accumulates it into the buffer.
func (r *accumulatedReader) Read(p []byte) (int, error) {
n, err := r.reader.Read(p)
if n > 0 {
r.buffer.Write(p[:n])
}
return n, err
}
func (r *accumulatedReader) Restore() {
r.reader = io.MultiReader(r.buffer, r.reader)
r.buffer = &bytes.Buffer{}
}
type putResponse struct {
ObjectID string `json:"object_id"`
ContainerID string `json:"container_id"`
@ -44,11 +77,7 @@ func (pr *putResponse) encode(w io.Writer) error {
// Upload handles multipart upload request.
func (h *Handler) Upload(c *fasthttp.RequestCtx) {
var (
file MultipartFile
idObj oid.ID
addr oid.Address
)
var file MultipartFile
scid, _ := c.UserValue("cid").(string)
bodyStream := c.RequestBodyStream()
@ -64,32 +93,62 @@ func (h *Handler) Upload(c *fasthttp.RequestCtx) {
return
}
defer func() {
// If the temporary reader can be closed - let's close it.
if file == nil {
return
}
err := file.Close()
log.Debug(
logs.CloseTemporaryMultipartFormFile,
zap.Stringer("address", addr),
zap.String("filename", file.FileName()),
zap.Error(err),
)
}()
boundary := string(c.Request.Header.MultipartFormBoundary())
if file, err = fetchMultipartFile(log, bodyStream, boundary); err != nil {
log.Error(logs.CouldNotReceiveMultipartForm, zap.Error(err))
response.Error(c, "could not receive multipart/form: "+err.Error(), fasthttp.StatusBadRequest)
return
}
defer func() {
if err := file.Close(); err != nil {
log.Warn(logs.FailedToCloseTemporaryMultipartFormFile, zap.Error(err))
}
}()
if c.Request.Header.Peek(explodeArchiveHeader) != nil {
h.explodeArchive(c, log, bktInfo, file)
} else {
h.uploadSingleObject(c, log, bktInfo, file)
}
// Multipart is multipart and thus can contain more than one part which
// we ignore at the moment. Also, when dealing with chunked encoding
// the last zero-length chunk might be left unread (because multipart
// reader only cares about its boundary and doesn't look further) and
// it will be (erroneously) interpreted as the start of the next
// pipelined header. Thus we need to drain the body buffer.
for {
_, err = bodyStream.Read(drainBuf)
if err == io.EOF || errors.Is(err, io.ErrUnexpectedEOF) {
break
}
}
}
func (h *Handler) uploadSingleObject(c *fasthttp.RequestCtx, log *zap.Logger, bktInfo *data.BucketInfo, file MultipartFile) {
idObj, err := h.uploadObject(c, log, bktInfo, file.FileName(), file)
if err != nil {
log.Error(logs.FailedToUploadObject, zap.Error(err))
return
}
addr := newAddress(bktInfo.CID, idObj)
c.Response.Header.SetContentType(jsonHeader)
// Try to return the response, otherwise, if something went wrong, throw an error.
if err = newPutResponse(addr).encode(c); err != nil {
log.Error(logs.CouldNotEncodeResponse, zap.Error(err))
response.Error(c, "could not encode response", fasthttp.StatusBadRequest)
return
}
}
func (h *Handler) uploadObject(c *fasthttp.RequestCtx, log *zap.Logger, bktInfo *data.BucketInfo, fileName string, file io.Reader) (oid.ID, error) {
ctx := utils.GetContextFromRequest(c)
filtered, err := filterHeaders(log, &c.Request.Header)
if err != nil {
log.Error(logs.CouldNotProcessHeaders, zap.Error(err))
response.Error(c, err.Error(), fasthttp.StatusBadRequest)
return
log.Error(logs.FailedToFilterHeaders, zap.Error(err))
response.Error(c, "could not filter headers", fasthttp.StatusBadRequest)
return oid.ID{}, err
}
now := time.Now()
@ -104,7 +163,7 @@ func (h *Handler) Upload(c *fasthttp.RequestCtx) {
if err = utils.PrepareExpirationHeader(c, h.frostfs, filtered, now); err != nil {
log.Error(logs.CouldNotPrepareExpirationHeader, zap.Error(err))
response.Error(c, "could not prepare expiration header: "+err.Error(), fasthttp.StatusBadRequest)
return
return oid.ID{}, err
}
attributes := make([]object.Attribute, 0, len(filtered))
@ -117,10 +176,10 @@ func (h *Handler) Upload(c *fasthttp.RequestCtx) {
}
// sets FileName attribute if it wasn't set from header
if _, ok := filtered[object.AttributeFileName]; !ok {
filename := object.NewAttribute()
filename.SetKey(object.AttributeFileName)
filename.SetValue(file.FileName())
attributes = append(attributes, *filename)
fileNameAttr := object.NewAttribute()
fileNameAttr.SetKey(object.AttributeFileName)
fileNameAttr.SetValue(fileName)
attributes = append(attributes, *fileNameAttr)
}
// sets Timestamp attribute if it wasn't set from header and enabled by settings
if _, ok := filtered[object.AttributeTimestamp]; !ok && h.config.DefaultTimestamp() {
@ -146,36 +205,59 @@ func (h *Handler) Upload(c *fasthttp.RequestCtx) {
BufferMaxSize: h.config.BufferMaxSizeForPut(),
}
var idObj oid.ID
if idObj, err = h.frostfs.CreateObject(ctx, prm); err != nil {
h.handlePutFrostFSErr(c, err, log)
return
return oid.ID{}, err
}
log.Debug(logs.ObjectUploaded,
zap.String("oid", idObj.EncodeToString()),
zap.String("FileName", fileName),
)
return idObj, nil
}
// explodeArchive read files from archive and creates objects for each of them.
// Sets FilePath attribute with name from tar.Header.
func (h *Handler) explodeArchive(c *fasthttp.RequestCtx, log *zap.Logger, bktInfo *data.BucketInfo, formFile io.ReadCloser) {
// default reader - without gzip decompression
accReader := newAccumulatedReader(formFile)
var reader io.Reader
if gzipReader, err := gzip.NewReader(accReader); err == nil {
reader = gzipReader
defer func() {
if err := gzipReader.Close(); err != nil {
log.Warn(logs.FailedToCloseReader, zap.Error(err))
}
}()
} else {
log.Info(logs.CompressionCheckFailed, zap.Error(err))
accReader.Restore()
reader = accReader
}
addr.SetObject(idObj)
addr.SetContainer(bktInfo.CID)
// Try to return the response, otherwise, if something went wrong, throw an error.
if err = newPutResponse(addr).encode(c); err != nil {
log.Error(logs.CouldNotEncodeResponse, zap.Error(err))
response.Error(c, "could not encode response", fasthttp.StatusBadRequest)
return
}
// Multipart is multipart and thus can contain more than one part which
// we ignore at the moment. Also, when dealing with chunked encoding
// the last zero-length chunk might be left unread (because multipart
// reader only cares about its boundary and doesn't look further) and
// it will be (erroneously) interpreted as the start of the next
// pipelined header. Thus we need to drain the body buffer.
tarReader := tar.NewReader(reader)
for {
_, err = bodyStream.Read(drainBuf)
if err == io.EOF || err == io.ErrUnexpectedEOF {
obj, err := tarReader.Next()
if errors.Is(err, io.EOF) {
break
} else if err != nil {
log.Error(logs.FailedToReadFileFromTar, zap.Error(err))
response.Error(c, "could not read tar header: "+err.Error(), fasthttp.StatusBadRequest)
return
}
if isDir(obj.Name) {
continue
}
c.Request.Header.Set(utils.UserAttributeHeaderPrefix+object.AttributeFilePath, obj.Name)
_, err = h.uploadObject(c, log, bktInfo, filepath.Base(obj.Name), tarReader)
if err != nil {
log.Error(logs.FailedToUploadObject, zap.Error(err))
response.Error(c, "could not upload object: "+err.Error(), fasthttp.StatusBadRequest)
return
}
}
// Report status code and content type.
c.Response.SetStatusCode(fasthttp.StatusOK)
c.Response.Header.SetContentType(jsonHeader)
}
func (h *Handler) handlePutFrostFSErr(r *fasthttp.RequestCtx, err error, log *zap.Logger) {

1
internal/handler/upload_test.go Normal file
View file

@ -0,0 +1 @@
package handler

8
internal/logs/logs.go
View file

@ -25,7 +25,7 @@ const (
CantGracefullyShutDownService = "can't gracefully shut down service, force stop" // Error in ../../metrics/service.go
IgnorePartEmptyFormName = "ignore part, empty form name" // Debug in ../../uploader/upload.go
IgnorePartEmptyFilename = "ignore part, empty filename" // Debug in ../../uploader/upload.go
CloseTemporaryMultipartFormFile = "close temporary multipart/form file" // Debug in ../../uploader/upload.go
FailedToCloseTemporaryMultipartFormFile = "failed to close temporary multipart/form file" // Warn in ../../uploader/upload.go
CouldNotReceiveMultipartForm = "could not receive multipart/form" // Error in ../../uploader/upload.go
CouldNotParseClientTime = "could not parse client time" // Warn in ../../uploader/upload.go
CouldNotPrepareExpirationHeader = "could not prepare expiration header" // Error in ../../uploader/upload.go
@ -89,4 +89,10 @@ const (
MultinetDialFail = "multinet dial failed"
FailedToLoadMultinetConfig = "failed to load multinet config"
MultinetConfigWontBeUpdated = "multinet config won't be updated"
FailedToFilterHeaders = "failed to filter headers"
FailedToReadFileFromTar = "failed to read file from tar"
FailedToUploadObject = "failed to upload object"
ObjectUploaded = "object uploaded"
FailedToCloseReader = "failed to close reader"
CompressionCheckFailed = "compression check failed"
)