3795cc2bd2
This paves the way for better object lifetime management. Concretely, it makes it possible to: - have different stores backing chain and state data. - having the same datastore library, but using different parameters. - attach different caching layers/policies to each class of data, e.g. sizing caches differently. - specifying different retention policies for chain and state data. This separation is important because: - access patterns/frequency of chain and state data are different. - state is derivable from chain, so one could never expunge the chain store, and only retain state objects reachable from the last finality in the state store.
261 lines
5.7 KiB
Go
261 lines
5.7 KiB
Go
package main
|
|
|
|
import (
|
|
"context"
|
|
"fmt"
|
|
"io"
|
|
|
|
"github.com/filecoin-project/go-state-types/abi"
|
|
"github.com/ipfs/bbloom"
|
|
"github.com/ipfs/go-cid"
|
|
"github.com/urfave/cli/v2"
|
|
"golang.org/x/xerrors"
|
|
|
|
badgerbs "github.com/filecoin-project/lotus/blockstore/badger"
|
|
"github.com/filecoin-project/lotus/chain/store"
|
|
"github.com/filecoin-project/lotus/chain/vm"
|
|
"github.com/filecoin-project/lotus/extern/sector-storage/ffiwrapper"
|
|
"github.com/filecoin-project/lotus/node/repo"
|
|
)
|
|
|
|
type cidSet interface {
|
|
Add(cid.Cid)
|
|
Has(cid.Cid) bool
|
|
HasRaw([]byte) bool
|
|
Len() int
|
|
}
|
|
|
|
type bloomSet struct {
|
|
bloom *bbloom.Bloom
|
|
}
|
|
|
|
func newBloomSet(size int64) (*bloomSet, error) {
|
|
b, err := bbloom.New(float64(size), 3)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
return &bloomSet{bloom: b}, nil
|
|
}
|
|
|
|
func (bs *bloomSet) Add(c cid.Cid) {
|
|
bs.bloom.Add(c.Hash())
|
|
|
|
}
|
|
|
|
func (bs *bloomSet) Has(c cid.Cid) bool {
|
|
return bs.bloom.Has(c.Hash())
|
|
}
|
|
|
|
func (bs *bloomSet) HasRaw(b []byte) bool {
|
|
return bs.bloom.Has(b)
|
|
}
|
|
|
|
func (bs *bloomSet) Len() int {
|
|
return int(bs.bloom.ElementsAdded())
|
|
}
|
|
|
|
type mapSet struct {
|
|
m map[string]struct{}
|
|
}
|
|
|
|
func newMapSet() *mapSet {
|
|
return &mapSet{m: make(map[string]struct{})}
|
|
}
|
|
|
|
func (bs *mapSet) Add(c cid.Cid) {
|
|
bs.m[string(c.Hash())] = struct{}{}
|
|
}
|
|
|
|
func (bs *mapSet) Has(c cid.Cid) bool {
|
|
_, ok := bs.m[string(c.Hash())]
|
|
return ok
|
|
}
|
|
|
|
func (bs *mapSet) HasRaw(b []byte) bool {
|
|
_, ok := bs.m[string(b)]
|
|
return ok
|
|
}
|
|
|
|
func (bs *mapSet) Len() int {
|
|
return len(bs.m)
|
|
}
|
|
|
|
var stateTreePruneCmd = &cli.Command{
|
|
Name: "state-prune",
|
|
Description: "Deletes old state root data from local chainstore",
|
|
Flags: []cli.Flag{
|
|
&cli.StringFlag{
|
|
Name: "repo",
|
|
Value: "~/.lotus",
|
|
},
|
|
&cli.Int64Flag{
|
|
Name: "keep-from-lookback",
|
|
Usage: "keep stateroots at or newer than the current height minus this lookback",
|
|
Value: 1800, // 2 x finality
|
|
},
|
|
&cli.IntFlag{
|
|
Name: "delete-up-to",
|
|
Usage: "delete up to the given number of objects (used to run a faster 'partial' sync)",
|
|
},
|
|
&cli.BoolFlag{
|
|
Name: "use-bloom-set",
|
|
Usage: "use a bloom filter for the 'good' set instead of a map, reduces memory usage but may not clean up as much",
|
|
},
|
|
&cli.BoolFlag{
|
|
Name: "dry-run",
|
|
Usage: "only enumerate the good set, don't do any deletions",
|
|
},
|
|
&cli.BoolFlag{
|
|
Name: "only-ds-gc",
|
|
Usage: "Only run datastore GC",
|
|
},
|
|
&cli.IntFlag{
|
|
Name: "gc-count",
|
|
Usage: "number of times to run gc",
|
|
Value: 20,
|
|
},
|
|
},
|
|
Action: func(cctx *cli.Context) error {
|
|
ctx := context.TODO()
|
|
|
|
fsrepo, err := repo.NewFS(cctx.String("repo"))
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
lkrepo, err := fsrepo.Lock(repo.FullNode)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
defer lkrepo.Close() //nolint:errcheck
|
|
|
|
bs, err := lkrepo.Blockstore(ctx, repo.UniversalBlockstore)
|
|
if err != nil {
|
|
return fmt.Errorf("failed to open blockstore: %w", err)
|
|
}
|
|
|
|
defer func() {
|
|
if c, ok := bs.(io.Closer); ok {
|
|
if err := c.Close(); err != nil {
|
|
log.Warnf("failed to close blockstore: %s", err)
|
|
}
|
|
}
|
|
}()
|
|
|
|
// After migrating to native blockstores, this has been made
|
|
// database-specific.
|
|
badgbs, ok := bs.(*badgerbs.Blockstore)
|
|
if !ok {
|
|
return fmt.Errorf("only badger blockstores are supported")
|
|
}
|
|
|
|
mds, err := lkrepo.Datastore(context.Background(), "/metadata")
|
|
if err != nil {
|
|
return err
|
|
}
|
|
defer mds.Close() //nolint:errcheck
|
|
|
|
const DiscardRatio = 0.2
|
|
if cctx.Bool("only-ds-gc") {
|
|
fmt.Println("running datastore gc....")
|
|
for i := 0; i < cctx.Int("gc-count"); i++ {
|
|
if err := badgbs.DB.RunValueLogGC(DiscardRatio); err != nil {
|
|
return xerrors.Errorf("datastore GC failed: %w", err)
|
|
}
|
|
}
|
|
fmt.Println("gc complete!")
|
|
return nil
|
|
}
|
|
|
|
cs := store.NewChainStore(bs, bs, mds, vm.Syscalls(ffiwrapper.ProofVerifier), nil)
|
|
defer cs.Close() //nolint:errcheck
|
|
|
|
if err := cs.Load(); err != nil {
|
|
return fmt.Errorf("loading chainstore: %w", err)
|
|
}
|
|
|
|
var goodSet cidSet
|
|
if cctx.Bool("use-bloom-set") {
|
|
bset, err := newBloomSet(10000000)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
goodSet = bset
|
|
} else {
|
|
goodSet = newMapSet()
|
|
}
|
|
|
|
ts := cs.GetHeaviestTipSet()
|
|
|
|
rrLb := abi.ChainEpoch(cctx.Int64("keep-from-lookback"))
|
|
|
|
if err := cs.WalkSnapshot(ctx, ts, rrLb, true, true, func(c cid.Cid) error {
|
|
if goodSet.Len()%20 == 0 {
|
|
fmt.Printf("\renumerating keep set: %d ", goodSet.Len())
|
|
}
|
|
goodSet.Add(c)
|
|
return nil
|
|
}); err != nil {
|
|
return fmt.Errorf("snapshot walk failed: %w", err)
|
|
}
|
|
|
|
fmt.Println()
|
|
fmt.Printf("Successfully marked keep set! (%d objects)\n", goodSet.Len())
|
|
|
|
if cctx.Bool("dry-run") {
|
|
return nil
|
|
}
|
|
|
|
b := badgbs.DB.NewWriteBatch()
|
|
defer b.Cancel()
|
|
|
|
markForRemoval := func(c cid.Cid) error {
|
|
return b.Delete(badgbs.StorageKey(nil, c))
|
|
}
|
|
|
|
keys, err := bs.AllKeysChan(context.Background())
|
|
if err != nil {
|
|
return xerrors.Errorf("failed to query blockstore: %w", err)
|
|
}
|
|
|
|
dupTo := cctx.Int("delete-up-to")
|
|
|
|
var deleteCount int
|
|
var goodHits int
|
|
for k := range keys {
|
|
if goodSet.HasRaw(k.Bytes()) {
|
|
goodHits++
|
|
continue
|
|
}
|
|
|
|
if err := markForRemoval(k); err != nil {
|
|
return fmt.Errorf("failed to remove cid %s: %w", k, err)
|
|
}
|
|
|
|
if deleteCount%20 == 0 {
|
|
fmt.Printf("\rdeleting %d objects (good hits: %d)... ", deleteCount, goodHits)
|
|
}
|
|
|
|
if dupTo != 0 && deleteCount > dupTo {
|
|
break
|
|
}
|
|
}
|
|
|
|
if err := b.Flush(); err != nil {
|
|
return xerrors.Errorf("failed to flush final batch delete: %w", err)
|
|
}
|
|
|
|
fmt.Println("running datastore gc....")
|
|
for i := 0; i < cctx.Int("gc-count"); i++ {
|
|
if err := badgbs.DB.RunValueLogGC(DiscardRatio); err != nil {
|
|
return xerrors.Errorf("datastore GC failed: %w", err)
|
|
}
|
|
}
|
|
fmt.Println("gc complete!")
|
|
|
|
return nil
|
|
},
|
|
}
|