lotus/extern/sector-storage/worker_local.go
2022-05-20 17:04:51 +08:00

885 lines
27 KiB
Go

package sectorstorage
import (
"context"
"encoding/json"
"io"
"os"
"reflect"
"runtime"
"strconv"
"sync"
"sync/atomic"
"time"
"github.com/elastic/go-sysinfo"
"github.com/google/uuid"
"github.com/hashicorp/go-multierror"
"github.com/ipfs/go-cid"
"golang.org/x/xerrors"
ffi "github.com/filecoin-project/filecoin-ffi"
"github.com/filecoin-project/go-state-types/abi"
"github.com/filecoin-project/go-statestore"
"github.com/filecoin-project/specs-actors/actors/runtime/proof"
"github.com/filecoin-project/specs-storage/storage"
"github.com/filecoin-project/lotus/extern/sector-storage/ffiwrapper"
"github.com/filecoin-project/lotus/extern/sector-storage/sealtasks"
"github.com/filecoin-project/lotus/extern/sector-storage/stores"
"github.com/filecoin-project/lotus/extern/sector-storage/storiface"
)
var pathTypes = []storiface.SectorFileType{storiface.FTUnsealed, storiface.FTSealed, storiface.FTCache, storiface.FTUpdate, storiface.FTUpdateCache}
type WorkerConfig struct {
TaskTypes []sealtasks.TaskType
NoSwap bool
// IgnoreResourceFiltering enables task distribution to happen on this
// worker regardless of its currently available resources. Used in testing
// with the local worker.
IgnoreResourceFiltering bool
MaxParallelChallengeReads int // 0 = no limit
ChallengeReadTimeout time.Duration // 0 = no timeout
}
// used do provide custom proofs impl (mostly used in testing)
type ExecutorFunc func() (ffiwrapper.Storage, error)
type EnvFunc func(string) (string, bool)
type LocalWorker struct {
storage stores.Store
localStore *stores.Local
sindex stores.SectorIndex
ret storiface.WorkerReturn
executor ExecutorFunc
noSwap bool
envLookup EnvFunc
// see equivalent field on WorkerConfig.
ignoreResources bool
ct *workerCallTracker
acceptTasks map[sealtasks.TaskType]struct{}
running sync.WaitGroup
taskLk sync.Mutex
challengeThrottle chan struct{}
challengeReadTimeout time.Duration
session uuid.UUID
testDisable int64
closing chan struct{}
}
func newLocalWorker(executor ExecutorFunc, wcfg WorkerConfig, envLookup EnvFunc, store stores.Store, local *stores.Local, sindex stores.SectorIndex, ret storiface.WorkerReturn, cst *statestore.StateStore) *LocalWorker {
acceptTasks := map[sealtasks.TaskType]struct{}{}
for _, taskType := range wcfg.TaskTypes {
acceptTasks[taskType] = struct{}{}
}
w := &LocalWorker{
storage: store,
localStore: local,
sindex: sindex,
ret: ret,
ct: &workerCallTracker{
st: cst,
},
acceptTasks: acceptTasks,
executor: executor,
noSwap: wcfg.NoSwap,
envLookup: envLookup,
ignoreResources: wcfg.IgnoreResourceFiltering,
challengeReadTimeout: wcfg.ChallengeReadTimeout,
session: uuid.New(),
closing: make(chan struct{}),
}
if wcfg.MaxParallelChallengeReads > 0 {
w.challengeThrottle = make(chan struct{}, wcfg.MaxParallelChallengeReads)
}
if w.executor == nil {
w.executor = w.ffiExec
}
unfinished, err := w.ct.unfinished()
if err != nil {
log.Errorf("reading unfinished tasks: %+v", err)
return w
}
go func() {
for _, call := range unfinished {
hostname, osErr := os.Hostname()
if osErr != nil {
log.Errorf("get hostname err: %+v", err)
hostname = ""
}
err := storiface.Err(storiface.ErrTempWorkerRestart, xerrors.Errorf("worker [Hostname: %s] restarted", hostname))
// TODO: Handle restarting PC1 once support is merged
if doReturn(context.TODO(), call.RetType, call.ID, ret, nil, err) {
if err := w.ct.onReturned(call.ID); err != nil {
log.Errorf("marking call as returned failed: %s: %+v", call.RetType, err)
}
}
}
}()
return w
}
func NewLocalWorker(wcfg WorkerConfig, store stores.Store, local *stores.Local, sindex stores.SectorIndex, ret storiface.WorkerReturn, cst *statestore.StateStore) *LocalWorker {
return newLocalWorker(nil, wcfg, os.LookupEnv, store, local, sindex, ret, cst)
}
type localWorkerPathProvider struct {
w *LocalWorker
op storiface.AcquireMode
}
func (l *localWorkerPathProvider) AcquireSector(ctx context.Context, sector storage.SectorRef, existing storiface.SectorFileType, allocate storiface.SectorFileType, sealing storiface.PathType) (storiface.SectorPaths, func(), error) {
paths, storageIDs, err := l.w.storage.AcquireSector(ctx, sector, existing, allocate, sealing, l.op)
if err != nil {
return storiface.SectorPaths{}, nil, err
}
releaseStorage, err := l.w.localStore.Reserve(ctx, sector, allocate, storageIDs, storiface.FSOverheadSeal)
if err != nil {
return storiface.SectorPaths{}, nil, xerrors.Errorf("reserving storage space: %w", err)
}
log.Debugf("acquired sector %d (e:%d; a:%d): %v", sector, existing, allocate, paths)
return paths, func() {
releaseStorage()
for _, fileType := range pathTypes {
if fileType&allocate == 0 {
continue
}
sid := storiface.PathByType(storageIDs, fileType)
if err := l.w.sindex.StorageDeclareSector(ctx, storiface.ID(sid), sector.ID, fileType, l.op == storiface.AcquireMove); err != nil {
log.Errorf("declare sector error: %+v", err)
}
}
}, nil
}
func (l *LocalWorker) ffiExec() (ffiwrapper.Storage, error) {
return ffiwrapper.New(&localWorkerPathProvider{w: l})
}
type ReturnType string
const (
DataCid ReturnType = "DataCid"
AddPiece ReturnType = "AddPiece"
SealPreCommit1 ReturnType = "SealPreCommit1"
SealPreCommit2 ReturnType = "SealPreCommit2"
SealCommit1 ReturnType = "SealCommit1"
SealCommit2 ReturnType = "SealCommit2"
FinalizeSector ReturnType = "FinalizeSector"
FinalizeReplicaUpdate ReturnType = "FinalizeReplicaUpdate"
ReplicaUpdate ReturnType = "ReplicaUpdate"
ProveReplicaUpdate1 ReturnType = "ProveReplicaUpdate1"
ProveReplicaUpdate2 ReturnType = "ProveReplicaUpdate2"
GenerateSectorKey ReturnType = "GenerateSectorKey"
ReleaseUnsealed ReturnType = "ReleaseUnsealed"
MoveStorage ReturnType = "MoveStorage"
UnsealPiece ReturnType = "UnsealPiece"
Fetch ReturnType = "Fetch"
)
// in: func(WorkerReturn, context.Context, CallID, err string)
// in: func(WorkerReturn, context.Context, CallID, ret T, err string)
func rfunc(in interface{}) func(context.Context, storiface.CallID, storiface.WorkerReturn, interface{}, *storiface.CallError) error {
rf := reflect.ValueOf(in)
ft := rf.Type()
withRet := ft.NumIn() == 5
return func(ctx context.Context, ci storiface.CallID, wr storiface.WorkerReturn, i interface{}, err *storiface.CallError) error {
rctx := reflect.ValueOf(ctx)
rwr := reflect.ValueOf(wr)
rerr := reflect.ValueOf(err)
rci := reflect.ValueOf(ci)
var ro []reflect.Value
if withRet {
ret := reflect.ValueOf(i)
if i == nil {
ret = reflect.Zero(rf.Type().In(3))
}
ro = rf.Call([]reflect.Value{rwr, rctx, rci, ret, rerr})
} else {
ro = rf.Call([]reflect.Value{rwr, rctx, rci, rerr})
}
if !ro[0].IsNil() {
return ro[0].Interface().(error)
}
return nil
}
}
var returnFunc = map[ReturnType]func(context.Context, storiface.CallID, storiface.WorkerReturn, interface{}, *storiface.CallError) error{
DataCid: rfunc(storiface.WorkerReturn.ReturnDataCid),
AddPiece: rfunc(storiface.WorkerReturn.ReturnAddPiece),
SealPreCommit1: rfunc(storiface.WorkerReturn.ReturnSealPreCommit1),
SealPreCommit2: rfunc(storiface.WorkerReturn.ReturnSealPreCommit2),
SealCommit1: rfunc(storiface.WorkerReturn.ReturnSealCommit1),
SealCommit2: rfunc(storiface.WorkerReturn.ReturnSealCommit2),
FinalizeSector: rfunc(storiface.WorkerReturn.ReturnFinalizeSector),
ReleaseUnsealed: rfunc(storiface.WorkerReturn.ReturnReleaseUnsealed),
ReplicaUpdate: rfunc(storiface.WorkerReturn.ReturnReplicaUpdate),
ProveReplicaUpdate1: rfunc(storiface.WorkerReturn.ReturnProveReplicaUpdate1),
ProveReplicaUpdate2: rfunc(storiface.WorkerReturn.ReturnProveReplicaUpdate2),
GenerateSectorKey: rfunc(storiface.WorkerReturn.ReturnGenerateSectorKeyFromData),
FinalizeReplicaUpdate: rfunc(storiface.WorkerReturn.ReturnFinalizeReplicaUpdate),
MoveStorage: rfunc(storiface.WorkerReturn.ReturnMoveStorage),
UnsealPiece: rfunc(storiface.WorkerReturn.ReturnUnsealPiece),
Fetch: rfunc(storiface.WorkerReturn.ReturnFetch),
}
func (l *LocalWorker) asyncCall(ctx context.Context, sector storage.SectorRef, rt ReturnType, work func(ctx context.Context, ci storiface.CallID) (interface{}, error)) (storiface.CallID, error) {
ci := storiface.CallID{
Sector: sector.ID,
ID: uuid.New(),
}
if err := l.ct.onStart(ci, rt); err != nil {
log.Errorf("tracking call (start): %+v", err)
}
l.running.Add(1)
go func() {
defer l.running.Done()
ctx := &wctx{
vals: ctx,
closing: l.closing,
}
res, err := work(ctx, ci)
if err != nil {
rb, err := json.Marshal(res)
if err != nil {
log.Errorf("tracking call (marshaling results): %+v", err)
} else {
if err := l.ct.onDone(ci, rb); err != nil {
log.Errorf("tracking call (done): %+v", err)
}
}
}
if err != nil {
hostname, osErr := os.Hostname()
if osErr != nil {
log.Errorf("get hostname err: %+v", err)
}
err = xerrors.Errorf("%w [Hostname: %s]", err.Error(), hostname)
}
if doReturn(ctx, rt, ci, l.ret, res, toCallError(err)) {
if err := l.ct.onReturned(ci); err != nil {
log.Errorf("tracking call (done): %+v", err)
}
}
}()
return ci, nil
}
func toCallError(err error) *storiface.CallError {
var serr *storiface.CallError
if err != nil && !xerrors.As(err, &serr) {
serr = storiface.Err(storiface.ErrUnknown, err)
}
return serr
}
// doReturn tries to send the result to manager, returns true if successful
func doReturn(ctx context.Context, rt ReturnType, ci storiface.CallID, ret storiface.WorkerReturn, res interface{}, rerr *storiface.CallError) bool {
for {
err := returnFunc[rt](ctx, ci, ret, res, rerr)
if err == nil {
break
}
log.Errorf("return error, will retry in 5s: %s: %+v", rt, err)
select {
case <-time.After(5 * time.Second):
case <-ctx.Done():
log.Errorf("failed to return results: %s", ctx.Err())
// fine to just return, worker is most likely shutting down, and
// we didn't mark the result as returned yet, so we'll try to
// re-submit it on restart
return false
}
}
return true
}
func (l *LocalWorker) NewSector(ctx context.Context, sector storage.SectorRef) error {
sb, err := l.executor()
if err != nil {
return err
}
return sb.NewSector(ctx, sector)
}
func (l *LocalWorker) DataCid(ctx context.Context, pieceSize abi.UnpaddedPieceSize, pieceData storage.Data) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, storage.NoSectorRef, DataCid, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
return sb.DataCid(ctx, pieceSize, pieceData)
})
}
func (l *LocalWorker) AddPiece(ctx context.Context, sector storage.SectorRef, epcs []abi.UnpaddedPieceSize, sz abi.UnpaddedPieceSize, r io.Reader) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, AddPiece, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
return sb.AddPiece(ctx, sector, epcs, sz, r)
})
}
func (l *LocalWorker) Fetch(ctx context.Context, sector storage.SectorRef, fileType storiface.SectorFileType, ptype storiface.PathType, am storiface.AcquireMode) (storiface.CallID, error) {
return l.asyncCall(ctx, sector, Fetch, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
_, done, err := (&localWorkerPathProvider{w: l, op: am}).AcquireSector(ctx, sector, fileType, storiface.FTNone, ptype)
if err == nil {
done()
}
return nil, err
})
}
func (l *LocalWorker) SealPreCommit1(ctx context.Context, sector storage.SectorRef, ticket abi.SealRandomness, pieces []abi.PieceInfo) (storiface.CallID, error) {
return l.asyncCall(ctx, sector, SealPreCommit1, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
{
// cleanup previous failed attempts if they exist
if err := l.storage.Remove(ctx, sector.ID, storiface.FTSealed, true, nil); err != nil {
return nil, xerrors.Errorf("cleaning up sealed data: %w", err)
}
if err := l.storage.Remove(ctx, sector.ID, storiface.FTCache, true, nil); err != nil {
return nil, xerrors.Errorf("cleaning up cache data: %w", err)
}
}
sb, err := l.executor()
if err != nil {
return nil, err
}
return sb.SealPreCommit1(ctx, sector, ticket, pieces)
})
}
func (l *LocalWorker) SealPreCommit2(ctx context.Context, sector storage.SectorRef, phase1Out storage.PreCommit1Out) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, SealPreCommit2, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
return sb.SealPreCommit2(ctx, sector, phase1Out)
})
}
func (l *LocalWorker) SealCommit1(ctx context.Context, sector storage.SectorRef, ticket abi.SealRandomness, seed abi.InteractiveSealRandomness, pieces []abi.PieceInfo, cids storage.SectorCids) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, SealCommit1, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
return sb.SealCommit1(ctx, sector, ticket, seed, pieces, cids)
})
}
func (l *LocalWorker) SealCommit2(ctx context.Context, sector storage.SectorRef, phase1Out storage.Commit1Out) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, SealCommit2, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
return sb.SealCommit2(ctx, sector, phase1Out)
})
}
func (l *LocalWorker) ReplicaUpdate(ctx context.Context, sector storage.SectorRef, pieces []abi.PieceInfo) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, ReplicaUpdate, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
sealerOut, err := sb.ReplicaUpdate(ctx, sector, pieces)
return sealerOut, err
})
}
func (l *LocalWorker) ProveReplicaUpdate1(ctx context.Context, sector storage.SectorRef, sectorKey, newSealed, newUnsealed cid.Cid) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, ProveReplicaUpdate1, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
return sb.ProveReplicaUpdate1(ctx, sector, sectorKey, newSealed, newUnsealed)
})
}
func (l *LocalWorker) ProveReplicaUpdate2(ctx context.Context, sector storage.SectorRef, sectorKey, newSealed, newUnsealed cid.Cid, vanillaProofs storage.ReplicaVanillaProofs) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, ProveReplicaUpdate2, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
return sb.ProveReplicaUpdate2(ctx, sector, sectorKey, newSealed, newUnsealed, vanillaProofs)
})
}
func (l *LocalWorker) GenerateSectorKeyFromData(ctx context.Context, sector storage.SectorRef, commD cid.Cid) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, GenerateSectorKey, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
return nil, sb.GenerateSectorKeyFromData(ctx, sector, commD)
})
}
func (l *LocalWorker) FinalizeSector(ctx context.Context, sector storage.SectorRef, keepUnsealed []storage.Range) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, FinalizeSector, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
if err := sb.FinalizeSector(ctx, sector, keepUnsealed); err != nil {
return nil, xerrors.Errorf("finalizing sector: %w", err)
}
if len(keepUnsealed) == 0 {
if err := l.storage.Remove(ctx, sector.ID, storiface.FTUnsealed, true, nil); err != nil {
return nil, xerrors.Errorf("removing unsealed data: %w", err)
}
}
return nil, err
})
}
func (l *LocalWorker) FinalizeReplicaUpdate(ctx context.Context, sector storage.SectorRef, keepUnsealed []storage.Range) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, FinalizeReplicaUpdate, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
if err := sb.FinalizeReplicaUpdate(ctx, sector, keepUnsealed); err != nil {
return nil, xerrors.Errorf("finalizing sector: %w", err)
}
if len(keepUnsealed) == 0 {
if err := l.storage.Remove(ctx, sector.ID, storiface.FTUnsealed, true, nil); err != nil {
return nil, xerrors.Errorf("removing unsealed data: %w", err)
}
}
return nil, err
})
}
func (l *LocalWorker) ReleaseUnsealed(ctx context.Context, sector storage.SectorRef, safeToFree []storage.Range) (storiface.CallID, error) {
return storiface.UndefCall, xerrors.Errorf("implement me")
}
func (l *LocalWorker) Remove(ctx context.Context, sector abi.SectorID) error {
var err error
if rerr := l.storage.Remove(ctx, sector, storiface.FTSealed, true, nil); rerr != nil {
err = multierror.Append(err, xerrors.Errorf("removing sector (sealed): %w", rerr))
}
if rerr := l.storage.Remove(ctx, sector, storiface.FTCache, true, nil); rerr != nil {
err = multierror.Append(err, xerrors.Errorf("removing sector (cache): %w", rerr))
}
if rerr := l.storage.Remove(ctx, sector, storiface.FTUnsealed, true, nil); rerr != nil {
err = multierror.Append(err, xerrors.Errorf("removing sector (unsealed): %w", rerr))
}
return err
}
func (l *LocalWorker) MoveStorage(ctx context.Context, sector storage.SectorRef, types storiface.SectorFileType) (storiface.CallID, error) {
return l.asyncCall(ctx, sector, MoveStorage, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
if err := l.storage.MoveStorage(ctx, sector, types); err != nil {
return nil, xerrors.Errorf("move to storage: %w", err)
}
for _, fileType := range storiface.PathTypes {
if fileType&types == 0 {
continue
}
if err := l.storage.RemoveCopies(ctx, sector.ID, fileType); err != nil {
return nil, xerrors.Errorf("rm copies (t:%s, s:%v): %w", fileType, sector, err)
}
}
return nil, nil
})
}
func (l *LocalWorker) UnsealPiece(ctx context.Context, sector storage.SectorRef, index storiface.UnpaddedByteIndex, size abi.UnpaddedPieceSize, randomness abi.SealRandomness, cid cid.Cid) (storiface.CallID, error) {
sb, err := l.executor()
if err != nil {
return storiface.UndefCall, err
}
return l.asyncCall(ctx, sector, UnsealPiece, func(ctx context.Context, ci storiface.CallID) (interface{}, error) {
log.Debugf("worker will unseal piece now, sector=%+v", sector.ID)
if err = sb.UnsealPiece(ctx, sector, index, size, randomness, cid); err != nil {
return nil, xerrors.Errorf("unsealing sector: %w", err)
}
if err = l.storage.RemoveCopies(ctx, sector.ID, storiface.FTSealed); err != nil {
return nil, xerrors.Errorf("removing source data: %w", err)
}
if err = l.storage.RemoveCopies(ctx, sector.ID, storiface.FTCache); err != nil {
return nil, xerrors.Errorf("removing source data: %w", err)
}
log.Debugf("worker has unsealed piece, sector=%+v", sector.ID)
return nil, nil
})
}
func (l *LocalWorker) GenerateWinningPoSt(ctx context.Context, ppt abi.RegisteredPoStProof, mid abi.ActorID, sectors []storiface.PostSectorChallenge, randomness abi.PoStRandomness) ([]proof.PoStProof, error) {
sb, err := l.executor()
if err != nil {
return nil, err
}
// don't throttle winningPoSt
// * Always want it done asap
// * It's usually just one sector
var wg sync.WaitGroup
wg.Add(len(sectors))
vproofs := make([][]byte, len(sectors))
var rerr error
for i, s := range sectors {
go func(i int, s storiface.PostSectorChallenge) {
defer wg.Done()
if l.challengeReadTimeout > 0 {
var cancel context.CancelFunc
ctx, cancel = context.WithTimeout(ctx, l.challengeReadTimeout)
defer cancel()
}
vanilla, err := l.storage.GenerateSingleVanillaProof(ctx, mid, s, ppt)
if err != nil {
rerr = multierror.Append(rerr, xerrors.Errorf("get winning sector:%d,vanila failed: %w", s.SectorNumber, err))
return
}
if vanilla == nil {
rerr = multierror.Append(rerr, xerrors.Errorf("get winning sector:%d,vanila is nil", s.SectorNumber))
}
vproofs[i] = vanilla
}(i, s)
}
wg.Wait()
if rerr != nil {
return nil, rerr
}
return sb.GenerateWinningPoStWithVanilla(ctx, ppt, mid, randomness, vproofs)
}
func (l *LocalWorker) GenerateWindowPoSt(ctx context.Context, ppt abi.RegisteredPoStProof, mid abi.ActorID, sectors []storiface.PostSectorChallenge, partitionIdx int, randomness abi.PoStRandomness) (storiface.WindowPoStResult, error) {
sb, err := l.executor()
if err != nil {
return storiface.WindowPoStResult{}, err
}
var slk sync.Mutex
var skipped []abi.SectorID
var wg sync.WaitGroup
wg.Add(len(sectors))
vproofs := make([][]byte, len(sectors))
for i, s := range sectors {
if l.challengeThrottle != nil {
select {
case l.challengeThrottle <- struct{}{}:
case <-ctx.Done():
return storiface.WindowPoStResult{}, xerrors.Errorf("context error waiting on challengeThrottle %w", err)
}
}
go func(i int, s storiface.PostSectorChallenge) {
defer wg.Done()
defer func() {
if l.challengeThrottle != nil {
<-l.challengeThrottle
}
}()
if l.challengeReadTimeout > 0 {
var cancel context.CancelFunc
ctx, cancel = context.WithTimeout(ctx, l.challengeReadTimeout)
defer cancel()
}
vanilla, err := l.storage.GenerateSingleVanillaProof(ctx, mid, s, ppt)
slk.Lock()
defer slk.Unlock()
if err != nil || vanilla == nil {
skipped = append(skipped, abi.SectorID{
Miner: mid,
Number: s.SectorNumber,
})
log.Errorf("reading PoSt challenge for sector %d, vlen:%d, err: %s", s.SectorNumber, len(vanilla), err)
return
}
vproofs[i] = vanilla
}(i, s)
}
wg.Wait()
if len(skipped) > 0 {
// This should happen rarely because before entering GenerateWindowPoSt we check all sectors by reading challenges.
// When it does happen, window post runner logic will just re-check sectors, and retry with newly-discovered-bad sectors skipped
log.Errorf("couldn't read some challenges (skipped %d)", len(skipped))
// note: can't return an error as this in an jsonrpc call
return storiface.WindowPoStResult{Skipped: skipped}, nil
}
res, err := sb.GenerateWindowPoStWithVanilla(ctx, ppt, mid, randomness, vproofs, partitionIdx)
return storiface.WindowPoStResult{
PoStProofs: res,
Skipped: skipped,
}, err
}
func (l *LocalWorker) TaskTypes(context.Context) (map[sealtasks.TaskType]struct{}, error) {
l.taskLk.Lock()
defer l.taskLk.Unlock()
return l.acceptTasks, nil
}
func (l *LocalWorker) TaskDisable(ctx context.Context, tt sealtasks.TaskType) error {
l.taskLk.Lock()
defer l.taskLk.Unlock()
delete(l.acceptTasks, tt)
return nil
}
func (l *LocalWorker) TaskEnable(ctx context.Context, tt sealtasks.TaskType) error {
l.taskLk.Lock()
defer l.taskLk.Unlock()
l.acceptTasks[tt] = struct{}{}
return nil
}
func (l *LocalWorker) Paths(ctx context.Context) ([]storiface.StoragePath, error) {
return l.localStore.Local(ctx)
}
func (l *LocalWorker) memInfo() (memPhysical, memUsed, memSwap, memSwapUsed uint64, err error) {
h, err := sysinfo.Host()
if err != nil {
return 0, 0, 0, 0, err
}
mem, err := h.Memory()
if err != nil {
return 0, 0, 0, 0, err
}
memPhysical = mem.Total
// mem.Available is memory available without swapping, it is more relevant for this calculation
memUsed = mem.Total - mem.Available
memSwap = mem.VirtualTotal
memSwapUsed = mem.VirtualUsed
if cgMemMax, cgMemUsed, cgSwapMax, cgSwapUsed, err := cgroupV1Mem(); err == nil {
if cgMemMax > 0 && cgMemMax < memPhysical {
memPhysical = cgMemMax
memUsed = cgMemUsed
}
if cgSwapMax > 0 && cgSwapMax < memSwap {
memSwap = cgSwapMax
memSwapUsed = cgSwapUsed
}
}
if cgMemMax, cgMemUsed, cgSwapMax, cgSwapUsed, err := cgroupV2Mem(); err == nil {
if cgMemMax > 0 && cgMemMax < memPhysical {
memPhysical = cgMemMax
memUsed = cgMemUsed
}
if cgSwapMax > 0 && cgSwapMax < memSwap {
memSwap = cgSwapMax
memSwapUsed = cgSwapUsed
}
}
if l.noSwap {
memSwap = 0
memSwapUsed = 0
}
return memPhysical, memUsed, memSwap, memSwapUsed, nil
}
func (l *LocalWorker) Info(context.Context) (storiface.WorkerInfo, error) {
hostname, err := os.Hostname() // TODO: allow overriding from config
if err != nil {
panic(err)
}
gpus, err := ffi.GetGPUDevices()
if err != nil {
log.Errorf("getting gpu devices failed: %+v", err)
}
memPhysical, memUsed, memSwap, memSwapUsed, err := l.memInfo()
if err != nil {
return storiface.WorkerInfo{}, xerrors.Errorf("getting memory info: %w", err)
}
resEnv, err := storiface.ParseResourceEnv(func(key, def string) (string, bool) {
return l.envLookup(key)
})
if err != nil {
return storiface.WorkerInfo{}, xerrors.Errorf("interpreting resource env vars: %w", err)
}
// parallel-p1-limit
p1Limit := -1
if limit, ok := os.LookupEnv("PARALLEL_P1_LIMIT"); ok {
li, err := strconv.Atoi(limit)
if err != nil {
log.Errorf("failed to parse PARALLEL_P1_LIMIT env var, default=-1")
} else {
p1Limit = li
}
}
taskLimits := make(map[sealtasks.TaskType]*storiface.LimitConfig)
taskLimits[sealtasks.TTPreCommit1] = &storiface.LimitConfig{
LimitCount: p1Limit,
RunCount: 0,
}
return storiface.WorkerInfo{
Hostname: hostname,
IgnoreResources: l.ignoreResources,
TaskLimits: taskLimits,
Resources: storiface.WorkerResources{
MemPhysical: memPhysical,
MemUsed: memUsed,
MemSwap: memSwap,
MemSwapUsed: memSwapUsed,
CPUs: uint64(runtime.NumCPU()),
GPUs: gpus,
Resources: resEnv,
},
}, nil
}
func (l *LocalWorker) Session(ctx context.Context) (uuid.UUID, error) {
if atomic.LoadInt64(&l.testDisable) == 1 {
return uuid.UUID{}, xerrors.Errorf("disabled")
}
select {
case <-l.closing:
return ClosedWorkerID, nil
default:
return l.session, nil
}
}
func (l *LocalWorker) Close() error {
close(l.closing)
return nil
}
// WaitQuiet blocks as long as there are tasks running
func (l *LocalWorker) WaitQuiet() {
l.running.Wait()
}
type wctx struct {
vals context.Context
closing chan struct{}
}
func (w *wctx) Deadline() (time.Time, bool) {
return time.Time{}, false
}
func (w *wctx) Done() <-chan struct{} {
return w.closing
}
func (w *wctx) Err() error {
select {
case <-w.closing:
return context.Canceled
default:
return nil
}
}
func (w *wctx) Value(key interface{}) interface{} {
return w.vals.Value(key)
}
var _ context.Context = &wctx{}
var _ Worker = &LocalWorker{}