Add transaction buffer filter.

filter/txnBuf/state.go

+// Copyright 2015 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+package txnBuf
+
+import (
+        "bytes"
+        "sync"
+
+        "github.com/luci/gae/impl/memory"
+        "github.com/luci/gae/service/datastore"
+        "github.com/luci/gae/service/datastore/serialize"
+        "github.com/luci/gae/service/info"
+        "github.com/luci/luci-go/common/errors"
+        "github.com/luci/luci-go/common/stringset"
+        "golang.org/x/net/context"
+)
+
+// DefaultSizeBudget is the size budget for the root transaction.
+//
+// Because our estimation algorithm isn't entirely correct, we take 5% off
+// the limit for encoding and estimate inaccuracies.
+//
+// 10MB taken on 2015/09/24:
+// https://cloud.google.com/appengine/docs/go/datastore/#Go_Quotas_and_limits
+const DefaultSizeBudget = int64((10 * 1000 * 1000) * 0.95)
+
+// DefaultSizeThreshold prevents the root transaction from getting too close
+// to the budget. If the code attempts to begin a transaction which would have
+// less than this threshold for its budget, the transaction will immediately
+// return ErrTransactionTooLarge.
+const DefaultSizeThreshold = int64(10 * 1000)
+
+// XGTransactionGroupLimit is the number of transaction groups to allow in an
+// XG transaction.
+//
+// 25 taken on 2015/09/24:
+// https://cloud.google.com/appengine/docs/go/datastore/transactions#Go_What_can_be_done_in_a_transaction
+const XGTransactionGroupLimit = 25
+
+type sizeTracker struct {
+        keyToSize map[string]int64
+        total     int64
+}
+
+func (s *sizeTracker) set(key string, val int64) {
+        prev, existed := s.keyToSize[key]
+        if s.keyToSize == nil {
+                s.keyToSize = make(map[string]int64)
+        }
+        s.keyToSize[key] = val
+        s.total += val - prev
+        if !existed {
+                s.total += int64(len(key))
+        }
+}
+
+func (s *sizeTracker) get(key string) (int64, bool) {
+        size, has := s.keyToSize[key]
+        return size, has
+}
+
+func (s *sizeTracker) has(key string) bool {
+        _, has := s.keyToSize[key]
+        return has
+}
+
+func (s *sizeTracker) dup() *sizeTracker {
+        if len(s.keyToSize) == 0 {
+                return &sizeTracker{}
+        }
+        k2s := make(map[string]int64, len(s.keyToSize))
+        for k, v := range s.keyToSize {
+                k2s[k] = v
+        }
+        return &sizeTracker{k2s, s.total}
+}
+
+type txnBufState struct {
+        sync.Mutex
+
+        // encoded key -> size of entity. A size of 0 means that the entity is
+        // deleted.
+        entState *sizeTracker
+        memDS    datastore.RawInterface
+
+        roots     stringset.Set
+        rootLimit int
+
+        aid         string
+        ns          string
+        parentDS    datastore.RawInterface
+        parentState *txnBufState
+
+        // sizeBudget is the number of bytes that this transaction has to operate
+        // within. It's only used when attempting to apply() the transaction, and
+        // it is the threshold for the delta of applying this transaction to the
+        // parent transaction. Note that a buffered transaction could actually have
+        // a negative delta if the parent transaction had many large entities which
+        // the inner transaction deleted.
+        sizeBudget int64
+
+        siblingLock sync.Mutex
+}
+
+func withTxnBuf(ctx context.Context, cb func(context.Context) error, opts *datastore.TransactionOptions) error {
+        inf := info.Get(ctx)
+        ns := inf.GetNamespace()
+
+        parentState, _ := ctx.Value(dsTxnBufParent).(*txnBufState)
+        roots := stringset.New(0)
+        rootLimit := 1
+        if opts != nil && opts.XG {
+                rootLimit = XGTransactionGroupLimit
+        }
+        sizeBudget := DefaultSizeBudget
+        if parentState != nil {
+                parentState.siblingLock.Lock()
+                defer parentState.siblingLock.Unlock()
+
+                // TODO(riannucci): this is a bit wonky since it means that a child
+                // transaction declaring XG=true will only get to modify 25 groups IF
+                // they're same groups affected by the parent transactions. So instead of
+                // respecting opts.XG for inner transactions, we just dup everything from
+                // the parent transaction.
+                roots = parentState.roots.Dup()
+                rootLimit = parentState.rootLimit
+
+                sizeBudget = parentState.sizeBudget - parentState.entState.total
+                if sizeBudget < DefaultSizeThreshold {
+                        return ErrTransactionTooLarge
+                }
+        }
+
+        memDS := memory.NewDatastore(ns)
+        {

[Vadim Sh., 2015/09/28 18:52:56]

Remove...

[iannucci, 2015/09/29 03:21:38]

k

+                t := memDS.Testable()
+                t.Consistent(true)
+                t.AutoIndex(true)
+                t.DisableSpecialEntities(true)
+        }
+        state := &txnBufState{
+                entState:    &sizeTracker{},
+                memDS:       memDS,
+                roots:       roots,
+                rootLimit:   rootLimit,
+                ns:          ns,
+                aid:         inf.AppID(),
+                parentDS:    datastore.Get(ctx).Raw(),
+                parentState: parentState,
+                sizeBudget:  sizeBudget,
+        }
+        ctx = context.WithValue(ctx, dsTxnBufParent, state)
+        err := cb(ctx)
+        if err != nil {
+                return err
+        }
+        return state.apply()
+}
+
+type item struct {
+        key      *datastore.Key
+        data     datastore.PropertyMap
+        buffered bool
+
+        encKey string
+        cmpRow string
+        err    error
+}
+
+func (i *item) getEncKey() string {
+        if i.encKey == "" {
+                i.encKey = string(serialize.ToBytes(i.key))
+        }
+        return i.encKey
+}
+
+func (i *item) getCmpRow(lower, upper []byte, order []datastore.IndexColumn) string {
+        if i.cmpRow == "" {
+                row, key := toComparableString(lower, upper, order, i.key, i.data)
+                i.cmpRow = string(row)
+                if i.encKey == "" {
+                        i.encKey = string(key)
+                }
+        }
+        return i.cmpRow
+}
+
+func (t *txnBufState) updateRootsLocked(roots stringset.Set) error {
+        curRootLen := t.roots.Len()
+        proposedRoots := stringset.New(1)
+        roots.Iter(func(root string) bool {
+                if !t.roots.Has(root) {
+                        proposedRoots.Add(root)
+                }
+                return proposedRoots.Len()+curRootLen <= t.rootLimit
+        })
+        if proposedRoots.Len()+curRootLen > t.rootLimit {
+                return errors.New("operating on too many entity groups in nested transaction")
+        }
+        // only need to update the roots if they did something that required updating
+        if proposedRoots.Len() > 0 {
+                proposedRoots.Iter(func(root string) bool {
+                        t.roots.Add(root)
+                        return true
+                })
+        }
+        return nil
+}
+
+func (t *txnBufState) getMulti(keys []*datastore.Key) ([]item, error) {
+        encKeys, roots := toEncoded(keys)
+        ret := make([]item, len(keys))
+
+        idxMap := []int(nil)
+        toGetKeys := []*datastore.Key(nil)
+
+        t.Lock()
+        defer t.Unlock()
+
+        if err := t.updateRootsLocked(roots); err != nil {
+                return nil, err
+        }
+
+        for i, key := range keys {
+                ret[i].key = key
+                ret[i].encKey = encKeys[i]
+                if size, ok := t.entState.get(ret[i].getEncKey()); ok {
+                        ret[i].buffered = true
+                        if size > 0 {
+                                idxMap = append(idxMap, i)
+                                toGetKeys = append(toGetKeys, key)
+                        }
+                }
+        }
+
+        if len(toGetKeys) > 0 {
+                j := 0
+                t.memDS.GetMulti(toGetKeys, nil, func(pm datastore.PropertyMap, err error) {
+                        idx := idxMap[j]
+                        impossible(err)

[Vadim Sh., 2015/09/28 18:52:56]

nit: move it up one line

[iannucci, 2015/09/29 03:21:38]

done

+                        ret[idx].data = pm
+                        j++
+                })
+        }
+
+        return ret, nil
+}
+
+func (t *txnBufState) deleteMulti(keys []*datastore.Key) error {
+        encKeys, roots := toEncoded(keys)
+
+        t.Lock()
+        defer t.Unlock()
+
+        if err := t.updateRootsLocked(roots); err != nil {
+                return err
+        }
+
+        i := 0
+        err := t.memDS.DeleteMulti(keys, func(err error) {
+                impossible(err)
+                t.entState.set(encKeys[i], 0)
+                i++
+        })
+        impossible(err)
+        return nil
+}
+
+func (t *txnBufState) putMulti(keys []*datastore.Key, vals []datastore.PropertyMap) error {
+        encKeys, roots := toEncoded(keys)
+
+        t.Lock()
+        defer t.Unlock()
+
+        if err := t.updateRootsLocked(roots); err != nil {
+                return err
+        }
+
+        i := 0
+        err := t.memDS.PutMulti(keys, vals, func(k *datastore.Key, err error) {
+                impossible(err)
+                t.entState.set(encKeys[i], vals[i].EstimateSize())
+                i++
+        })
+        impossible(err)
+        return nil
+}
+
+func (t *txnBufState) apply() error {
+        t.Lock()
+        defer t.Unlock()
+
+        // if parentState is nil... just try to commit this anyway. The estimates
+        // we're using here are just educated guesses. If it fits for real, then
+        // hooray. If not, then the underlying datastore will error.
+        if t.parentState != nil {
+                proposedState := t.parentState.entState.dup()
+                for k, v := range t.entState.keyToSize {
+                        proposedState.set(k, v)
+                }
+                if proposedState.total > t.sizeBudget {
+                        return ErrTransactionTooLarge
+                }
+        }
+
+        toPutKeys := []*datastore.Key(nil)
+        toPut := []datastore.PropertyMap(nil)
+        toDel := []*datastore.Key(nil)
+
+        fq, err := datastore.NewQuery("").Finalize()

[Vadim Sh., 2015/09/28 18:52:56]

wat?

[iannucci, 2015/09/29 03:21:38]

added comment

+        impossible(err)
+
+        err = t.memDS.Run(fq, func(key *datastore.Key, data datastore.PropertyMap, _ datastore.CursorCB) bool {
+                toPutKeys = append(toPutKeys, key)
+                toPut = append(toPut, data)
+                return true
+        })
+        memoryCorruption(err)
+
+        for keyStr, size := range t.entState.keyToSize {
+                if size == 0 {
+                        k, err := serialize.ReadKey(bytes.NewBufferString(keyStr), serialize.WithoutContext, t.aid, t.ns)
+                        memoryCorruption(err)
+                        toDel = append(toDel, k)
+                }
+        }
+
+        wg := sync.WaitGroup{}
+
+        pErr := error(nil)
+        dErr := error(nil)
+
+        ds := t.parentDS
+        if toPut != nil {
+                wg.Add(1)
+                go func() {
+                        defer wg.Done()
+                        mErr := errors.NewLazyMultiError(len(toPut))
+                        i := 0
+                        pErr = ds.PutMulti(toPutKeys, toPut, func(_ *datastore.Key, err error) {
+                                i++
+                                mErr.Assign(i, err)
+                        })
+                        pErr = mErr.Get()
+                }()
+        }
+
+        if toDel != nil {
+                wg.Add(1)
+                go func() {
+                        defer wg.Done()
+                        mErr := errors.NewLazyMultiError(len(toDel))
+                        i := 0
+                        dErr = ds.DeleteMulti(toDel, func(err error) {
+                                mErr.Assign(i, err)
+                                i++
+                        })
+                        dErr = mErr.Get()
+                }()
+        }
+        wg.Wait()
+
+        if pErr != nil {
+                return pErr
+        }
+        return dErr
+}
+
+func toEncoded(keys []*datastore.Key) (full []string, roots stringset.Set) {
+        roots = stringset.New(len(keys))
+        full = make([]string, len(keys))
+        for i, k := range keys {
+                roots.Add(string(serialize.ToBytes(k.Root())))
+                full[i] = string(serialize.ToBytes(k))
+        }
+        return
+}