volume: automatically trim out unreachable entries

This commit is contained in:
Chris Lu 2020-10-28 01:14:39 -07:00
parent 3e925faddd
commit 744c5594da
2 changed files with 53 additions and 22 deletions

View file

@ -4,6 +4,7 @@ import (
"fmt"
"os"
"github.com/chrislusf/seaweedfs/weed/glog"
"github.com/chrislusf/seaweedfs/weed/storage/backend"
"github.com/chrislusf/seaweedfs/weed/storage/idx"
"github.com/chrislusf/seaweedfs/weed/storage/needle"
@ -11,17 +12,28 @@ import (
"github.com/chrislusf/seaweedfs/weed/util"
)
func CheckVolumeDataIntegrity(v *Volume, indexFile *os.File) (lastAppendAtNs uint64, e error) {
func CheckAndFixVolumeDataIntegrity(v *Volume, indexFile *os.File) (lastAppendAtNs uint64, err error) {
var indexSize int64
if indexSize, e = verifyIndexFileIntegrity(indexFile); e != nil {
return 0, fmt.Errorf("verifyIndexFileIntegrity %s failed: %v", indexFile.Name(), e)
if indexSize, err = verifyIndexFileIntegrity(indexFile); err != nil {
return 0, fmt.Errorf("verifyIndexFileIntegrity %s failed: %v", indexFile.Name(), err)
}
if indexSize == 0 {
return 0, nil
}
for i := 1; i <= 10; i++ {
// check and fix last 10 entries
lastAppendAtNs, err = doCheckAndFixVolumeData(v, indexFile, indexSize-int64(i)*NeedleMapEntrySize)
if err != ErrorSizeMismatch {
break
}
}
return
}
func doCheckAndFixVolumeData(v *Volume, indexFile *os.File, indexOffset int64) (lastAppendAtNs uint64, err error) {
var lastIdxEntry []byte
if lastIdxEntry, e = readIndexEntryAtOffset(indexFile, indexSize-NeedleMapEntrySize); e != nil {
return 0, fmt.Errorf("readLastIndexEntry %s failed: %v", indexFile.Name(), e)
if lastIdxEntry, err = readIndexEntryAtOffset(indexFile, indexOffset); err != nil {
return 0, fmt.Errorf("readLastIndexEntry %s failed: %v", indexFile.Name(), err)
}
key, offset, size := idx.IdxFileEntry(lastIdxEntry)
if offset.IsZero() {
@ -29,15 +41,15 @@ func CheckVolumeDataIntegrity(v *Volume, indexFile *os.File) (lastAppendAtNs uin
}
if size < 0 {
// read the deletion entry
if lastAppendAtNs, e = verifyDeletedNeedleIntegrity(v.DataBackend, v.Version(), key); e != nil {
return lastAppendAtNs, fmt.Errorf("verifyNeedleIntegrity %s failed: %v", indexFile.Name(), e)
if lastAppendAtNs, err = verifyDeletedNeedleIntegrity(v.DataBackend, v.Version(), key); err != nil {
return lastAppendAtNs, fmt.Errorf("verifyNeedleIntegrity %s failed: %v", indexFile.Name(), err)
}
} else {
if lastAppendAtNs, e = verifyNeedleIntegrity(v.DataBackend, v.Version(), offset.ToAcutalOffset(), key, size); e != nil {
return lastAppendAtNs, fmt.Errorf("verifyNeedleIntegrity %s failed: %v", indexFile.Name(), e)
if lastAppendAtNs, err = verifyNeedleIntegrity(v.DataBackend, v.Version(), offset.ToAcutalOffset(), key, size); err != nil {
return lastAppendAtNs, err
}
}
return
return lastAppendAtNs, nil
}
func verifyIndexFileIntegrity(indexFile *os.File) (indexSize int64, err error) {
@ -60,18 +72,37 @@ func readIndexEntryAtOffset(indexFile *os.File, offset int64) (bytes []byte, err
}
func verifyNeedleIntegrity(datFile backend.BackendStorageFile, v needle.Version, offset int64, key NeedleId, size Size) (lastAppendAtNs uint64, err error) {
n := new(needle.Needle)
// case: node total memory 8g, set volumeLimitSize=2048 , save 10 files, every file size 2.2g or more , when we restart the volume server , while see out of memory error
// fix: When the size of the last file exceeds 10M, consider directly returning the last modify time
if size > 10 * 1024 * 1024 {
bytes , err := needle.ReadNeedleBlob(datFile, offset+int64(size), 0, v);
if err == nil {
if v == needle.Version3 {
tsOffset := NeedleHeaderSize + 0 + needle.NeedleChecksumSize
n.AppendAtNs = util.BytesToUint64(bytes[tsOffset : tsOffset+TimestampSize])
}
n, _, _, err := needle.ReadNeedleHeader(datFile, v, offset)
if err != nil {
return 0, fmt.Errorf("read %s at %d", datFile.Name(), offset)
}
if n.Size != size {
return 0, ErrorSizeMismatch
}
if v == needle.Version3 {
bytes := make([]byte, TimestampSize)
_, err = datFile.ReadAt(bytes, offset+NeedleHeaderSize+int64(size)+needle.NeedleChecksumSize)
if err != nil {
return 0, fmt.Errorf("check %s entry offset %d size %d: %v", datFile.Name(), offset, size, err)
}
return n.AppendAtNs, err
n.AppendAtNs = util.BytesToUint64(bytes)
fileTailOffset := offset + needle.GetActualSize(size, v)
fileSize, _, err := datFile.GetStat()
if err != nil {
return 0, fmt.Errorf("stat file %s: %v", datFile.Name(), err)
}
if fileSize == fileTailOffset {
return n.AppendAtNs, nil
}
if fileSize > fileTailOffset {
glog.Warningf("Truncate %s from %d bytes to %d bytes!", datFile.Name(), fileSize, fileTailOffset)
err = datFile.Truncate(fileTailOffset)
if err == nil {
return n.AppendAtNs, nil
}
return n.AppendAtNs, fmt.Errorf("truncate file %s: %v", datFile.Name(), err)
}
glog.Warningf("data file %s has %d bytes, less than expected %d bytes!", datFile.Name(), fileSize, fileTailOffset)
}
if err = n.ReadData(datFile, offset, size, v); err != nil {
return n.AppendAtNs, fmt.Errorf("read data [%d,%d) : %v", offset, offset+int64(size), err)

View file

@ -89,7 +89,7 @@ func (v *Volume) load(alsoLoadIndex bool, createDatIfMissing bool, needleMapKind
return fmt.Errorf("cannot write Volume Index %s.idx: %v", fileName, err)
}
}
if v.lastAppendAtNs, err = CheckVolumeDataIntegrity(v, indexFile); err != nil {
if v.lastAppendAtNs, err = CheckAndFixVolumeDataIntegrity(v, indexFile); err != nil {
v.noWriteOrDelete = true
glog.V(0).Infof("volumeDataIntegrityChecking failed %v", err)
}