implement putMetacacheObject() optimizing List operations (#12903)

removes unexpected features from regular putObject() such as

- increasing parity when disks are down, avoids
  a lot of DiskInfo() calls.

- triggering MRF for metacache objects
  if disks are offline

- avoiding renames from temporary location
  to actual namespace, not needed since
  metacache files are unique.
This commit is contained in:
Harshavardhana 2021-08-09 06:58:54 -07:00 committed by GitHub
parent 92c94011f1
commit 54ab3a1d5b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 162 additions and 2 deletions

View File

@ -582,6 +582,167 @@ func rename(ctx context.Context, disks []StorageAPI, srcBucket, srcEntry, dstBuc
return evalDisks(disks, errs), err
}
func (er erasureObjects) putMetacacheObject(ctx context.Context, key string, r *PutObjReader, opts ObjectOptions) (objInfo ObjectInfo, err error) {
data := r.Reader
// No metadata is set, allocate a new one.
if opts.UserDefined == nil {
opts.UserDefined = make(map[string]string)
}
storageDisks := er.getDisks()
// Get parity and data drive count based on storage class metadata
parityDrives := globalStorageClass.GetParityForSC(opts.UserDefined[xhttp.AmzStorageClass])
if parityDrives <= 0 {
parityDrives = er.defaultParityCount
}
dataDrives := len(storageDisks) - parityDrives
// we now know the number of blocks this object needs for data and parity.
// writeQuorum is dataBlocks + 1
writeQuorum := dataDrives
if dataDrives == parityDrives {
writeQuorum++
}
// Validate input data size and it can never be less than zero.
if data.Size() < -1 {
logger.LogIf(ctx, errInvalidArgument, logger.Application)
return ObjectInfo{}, toObjectErr(errInvalidArgument)
}
// Initialize parts metadata
partsMetadata := make([]FileInfo, len(storageDisks))
fi := newFileInfo(pathJoin(minioMetaBucket, key), dataDrives, parityDrives)
fi.DataDir = mustGetUUID()
// Initialize erasure metadata.
for index := range partsMetadata {
partsMetadata[index] = fi
}
// Order disks according to erasure distribution
var onlineDisks []StorageAPI
onlineDisks, partsMetadata = shuffleDisksAndPartsMetadata(storageDisks, partsMetadata, fi)
erasure, err := NewErasure(ctx, fi.Erasure.DataBlocks, fi.Erasure.ParityBlocks, fi.Erasure.BlockSize)
if err != nil {
return ObjectInfo{}, toObjectErr(err, minioMetaBucket, key)
}
// Fetch buffer for I/O, returns from the pool if not allocates a new one and returns.
var buffer []byte
switch size := data.Size(); {
case size == 0:
buffer = make([]byte, 1) // Allocate atleast a byte to reach EOF
case size >= fi.Erasure.BlockSize:
buffer = er.bp.Get()
defer er.bp.Put(buffer)
case size < fi.Erasure.BlockSize:
// No need to allocate fully blockSizeV1 buffer if the incoming data is smaller.
buffer = make([]byte, size, 2*size+int64(fi.Erasure.ParityBlocks+fi.Erasure.DataBlocks-1))
}
if len(buffer) > int(fi.Erasure.BlockSize) {
buffer = buffer[:fi.Erasure.BlockSize]
}
shardFileSize := erasure.ShardFileSize(data.Size())
writers := make([]io.Writer, len(onlineDisks))
inlineBuffers := make([]*bytes.Buffer, len(onlineDisks))
for i, disk := range onlineDisks {
if disk == nil {
continue
}
inlineBuffers[i] = bytes.NewBuffer(make([]byte, 0, shardFileSize))
writers[i] = newStreamingBitrotWriterBuffer(inlineBuffers[i], DefaultBitrotAlgorithm, erasure.ShardSize())
}
n, erasureErr := erasure.Encode(ctx, data, writers, buffer, writeQuorum)
closeBitrotWriters(writers)
if erasureErr != nil {
return ObjectInfo{}, toObjectErr(erasureErr, minioMetaBucket, key)
}
// Should return IncompleteBody{} error when reader has fewer bytes
// than specified in request header.
if n < data.Size() {
return ObjectInfo{}, IncompleteBody{Bucket: minioMetaBucket, Object: key}
}
for i, w := range writers {
if w == nil {
onlineDisks[i] = nil
continue
}
if len(inlineBuffers) > 0 && inlineBuffers[i] != nil {
partsMetadata[i].Data = inlineBuffers[i].Bytes()
} else {
partsMetadata[i].Data = nil
}
partsMetadata[i].AddObjectPart(1, "", n, data.ActualSize())
partsMetadata[i].Erasure.AddChecksumInfo(ChecksumInfo{
PartNumber: 1,
Algorithm: DefaultBitrotAlgorithm,
Hash: bitrotWriterSum(w),
})
}
modTime := UTCNow()
// Fill all the necessary metadata.
// Update `xl.meta` content on each disks.
for index := range partsMetadata {
partsMetadata[index].Metadata = opts.UserDefined
partsMetadata[index].Size = n
partsMetadata[index].ModTime = modTime
}
// Set an additional header when data is inlined.
for index := range partsMetadata {
partsMetadata[index].SetInlineData()
}
for i := 0; i < len(onlineDisks); i++ {
if onlineDisks[i] != nil && onlineDisks[i].IsOnline() {
// Object info is the same in all disks, so we can pick
// the first meta from online disk
fi = partsMetadata[i]
break
}
}
g := errgroup.WithNErrs(len(onlineDisks))
// Rename file on all underlying storage disks.
for index := range onlineDisks {
index := index
g.Go(func() error {
if onlineDisks[index] == nil {
return errDiskNotFound
}
// Pick one FileInfo for a disk at index.
fi := partsMetadata[index]
// Assign index when index is initialized
if fi.Erasure.Index == 0 {
fi.Erasure.Index = index + 1
}
if fi.IsValid() {
return onlineDisks[index].WriteMetadata(ctx, minioMetaBucket, key, fi)
}
return errFileCorrupt
}, index)
}
// Wait for all renames to finish.
errs := g.Wait()
return fi.ToObjectInfo(minioMetaBucket, key), reduceWriteQuorumErrs(ctx, errs, objectOpIgnoredErrs, writeQuorum)
}
// PutObject - creates an object upon reading from the input stream
// until EOF, erasure codes the data across all disk and additionally
// writes `xl.meta` which carries the necessary metadata for future

View File

@ -648,9 +648,8 @@ func (er *erasureObjects) saveMetaCacheStream(ctx context.Context, mc *metaCache
r, err := hash.NewReader(bytes.NewReader(b.data), int64(len(b.data)), "", "", int64(len(b.data)))
logger.LogIf(ctx, err)
custom := b.headerKV()
_, err = er.putObject(ctx, minioMetaBucket, o.objectPath(b.n), NewPutObjReader(r), ObjectOptions{
_, err = er.putMetacacheObject(ctx, o.objectPath(b.n), NewPutObjReader(r), ObjectOptions{
UserDefined: custom,
NoLock: true, // No need to hold namespace lock, each prefix caches uniquely.
})
if err != nil {
mc.setErr(err.Error())