2
2
mirror of https://github.com/octoleo/restic.git synced 2024-11-27 15:26:37 +00:00
restic/internal/repository/repack.go
Michael Eischer 367449dede prune: Reduce memory allocations while repacking
The slicing operator `slice[low:high]` default to 0 for the lower bound and
len(slice) for the upper bound when either or both are not specified.
Fix the code to use `cap(slice)` to check for the slice capacity.
2020-08-16 11:34:01 +02:00

118 lines
3.1 KiB
Go

package repository
import (
"context"
"github.com/restic/restic/internal/debug"
"github.com/restic/restic/internal/errors"
"github.com/restic/restic/internal/fs"
"github.com/restic/restic/internal/pack"
"github.com/restic/restic/internal/restic"
)
// Repack takes a list of packs together with a list of blobs contained in
// these packs. Each pack is loaded and the blobs listed in keepBlobs is saved
// into a new pack. Returned is the list of obsolete packs which can then
// be removed.
func Repack(ctx context.Context, repo restic.Repository, packs restic.IDSet, keepBlobs restic.BlobSet, p *restic.Progress) (obsoletePacks restic.IDSet, err error) {
if p != nil {
p.Start()
defer p.Done()
}
debug.Log("repacking %d packs while keeping %d blobs", len(packs), len(keepBlobs))
for packID := range packs {
// load the complete pack into a temp file
h := restic.Handle{Type: restic.PackFile, Name: packID.String()}
tempfile, hash, packLength, err := DownloadAndHash(ctx, repo.Backend(), h)
if err != nil {
return nil, errors.Wrap(err, "Repack")
}
debug.Log("pack %v loaded (%d bytes), hash %v", packID, packLength, hash)
if !packID.Equal(hash) {
return nil, errors.Errorf("hash does not match id: want %v, got %v", packID, hash)
}
_, err = tempfile.Seek(0, 0)
if err != nil {
return nil, errors.Wrap(err, "Seek")
}
blobs, err := pack.List(repo.Key(), tempfile, packLength)
if err != nil {
return nil, err
}
debug.Log("processing pack %v, blobs: %v", packID, len(blobs))
var buf []byte
for _, entry := range blobs {
h := restic.BlobHandle{ID: entry.ID, Type: entry.Type}
if !keepBlobs.Has(h) {
continue
}
debug.Log(" process blob %v", h)
if uint(cap(buf)) < entry.Length {
buf = make([]byte, entry.Length)
}
buf = buf[:entry.Length]
n, err := tempfile.ReadAt(buf, int64(entry.Offset))
if err != nil {
return nil, errors.Wrap(err, "ReadAt")
}
if n != len(buf) {
return nil, errors.Errorf("read blob %v from %v: not enough bytes read, want %v, got %v",
h, tempfile.Name(), len(buf), n)
}
nonce, ciphertext := buf[:repo.Key().NonceSize()], buf[repo.Key().NonceSize():]
plaintext, err := repo.Key().Open(ciphertext[:0], nonce, ciphertext, nil)
if err != nil {
return nil, err
}
id := restic.Hash(plaintext)
if !id.Equal(entry.ID) {
debug.Log("read blob %v/%v from %v: wrong data returned, hash is %v",
h.Type, h.ID, tempfile.Name(), id)
return nil, errors.Errorf("read blob %v from %v: wrong data returned, hash is %v",
h, tempfile.Name(), id)
}
// We do want to save already saved blobs!
_, _, err = repo.SaveBlob(ctx, entry.Type, plaintext, entry.ID, true)
if err != nil {
return nil, err
}
debug.Log(" saved blob %v", entry.ID)
keepBlobs.Delete(h)
}
if err = tempfile.Close(); err != nil {
return nil, errors.Wrap(err, "Close")
}
if err = fs.RemoveIfExists(tempfile.Name()); err != nil {
return nil, errors.Wrap(err, "Remove")
}
if p != nil {
p.Report(restic.Stat{Blobs: 1})
}
}
if err := repo.Flush(ctx); err != nil {
return nil, err
}
return packs, nil
}