groupcache/groupcache.go
Tipp Moseley d995f8a668 Fixes a concurrency error in groupcache. Previously, multiple Get
requests could simultaneously result in a load().  Only requests that
enter singleflight Do concurrently would be deduped, so it was possible for
populateCache to be called multiple times for the same key.  That
would result in overcounting nbytes, and eventually leading to a
livelock where nbytes > cacheBytes, but there were no entries in the
cache.

Change-Id: I5b304ce82041c1310c31b662486744e86509cc53
2016-01-27 11:33:35 -08:00

490 lines
13 KiB
Go

/*
Copyright 2012 Google Inc.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
// Package groupcache provides a data loading mechanism with caching
// and de-duplication that works across a set of peer processes.
//
// Each data Get first consults its local cache, otherwise delegates
// to the requested key's canonical owner, which then checks its cache
// or finally gets the data. In the common case, many concurrent
// cache misses across a set of peers for the same key result in just
// one cache fill.
package groupcache
import (
"errors"
"math/rand"
"strconv"
"sync"
"sync/atomic"
pb "github.com/golang/groupcache/groupcachepb"
"github.com/golang/groupcache/lru"
"github.com/golang/groupcache/singleflight"
)
// A Getter loads data for a key.
type Getter interface {
// Get returns the value identified by key, populating dest.
//
// The returned data must be unversioned. That is, key must
// uniquely describe the loaded data, without an implicit
// current time, and without relying on cache expiration
// mechanisms.
Get(ctx Context, key string, dest Sink) error
}
// A GetterFunc implements Getter with a function.
type GetterFunc func(ctx Context, key string, dest Sink) error
func (f GetterFunc) Get(ctx Context, key string, dest Sink) error {
return f(ctx, key, dest)
}
var (
mu sync.RWMutex
groups = make(map[string]*Group)
initPeerServerOnce sync.Once
initPeerServer func()
)
// GetGroup returns the named group previously created with NewGroup, or
// nil if there's no such group.
func GetGroup(name string) *Group {
mu.RLock()
g := groups[name]
mu.RUnlock()
return g
}
// NewGroup creates a coordinated group-aware Getter from a Getter.
//
// The returned Getter tries (but does not guarantee) to run only one
// Get call at once for a given key across an entire set of peer
// processes. Concurrent callers both in the local process and in
// other processes receive copies of the answer once the original Get
// completes.
//
// The group name must be unique for each getter.
func NewGroup(name string, cacheBytes int64, getter Getter) *Group {
return newGroup(name, cacheBytes, getter, nil)
}
// If peers is nil, the peerPicker is called via a sync.Once to initialize it.
func newGroup(name string, cacheBytes int64, getter Getter, peers PeerPicker) *Group {
if getter == nil {
panic("nil Getter")
}
mu.Lock()
defer mu.Unlock()
initPeerServerOnce.Do(callInitPeerServer)
if _, dup := groups[name]; dup {
panic("duplicate registration of group " + name)
}
g := &Group{
name: name,
getter: getter,
peers: peers,
cacheBytes: cacheBytes,
loadGroup: &singleflight.Group{},
}
if fn := newGroupHook; fn != nil {
fn(g)
}
groups[name] = g
return g
}
// newGroupHook, if non-nil, is called right after a new group is created.
var newGroupHook func(*Group)
// RegisterNewGroupHook registers a hook that is run each time
// a group is created.
func RegisterNewGroupHook(fn func(*Group)) {
if newGroupHook != nil {
panic("RegisterNewGroupHook called more than once")
}
newGroupHook = fn
}
// RegisterServerStart registers a hook that is run when the first
// group is created.
func RegisterServerStart(fn func()) {
if initPeerServer != nil {
panic("RegisterServerStart called more than once")
}
initPeerServer = fn
}
func callInitPeerServer() {
if initPeerServer != nil {
initPeerServer()
}
}
// A Group is a cache namespace and associated data loaded spread over
// a group of 1 or more machines.
type Group struct {
name string
getter Getter
peersOnce sync.Once
peers PeerPicker
cacheBytes int64 // limit for sum of mainCache and hotCache size
// mainCache is a cache of the keys for which this process
// (amongst its peers) is authorative. That is, this cache
// contains keys which consistent hash on to this process's
// peer number.
mainCache cache
// hotCache contains keys/values for which this peer is not
// authorative (otherwise they would be in mainCache), but
// are popular enough to warrant mirroring in this process to
// avoid going over the network to fetch from a peer. Having
// a hotCache avoids network hotspotting, where a peer's
// network card could become the bottleneck on a popular key.
// This cache is used sparingly to maximize the total number
// of key/value pairs that can be stored globally.
hotCache cache
// loadGroup ensures that each key is only fetched once
// (either locally or remotely), regardless of the number of
// concurrent callers.
loadGroup flightGroup
// Stats are statistics on the group.
Stats Stats
}
// flightGroup is defined as an interface which flightgroup.Group
// satisfies. We define this so that we may test with an alternate
// implementation.
type flightGroup interface {
// Done is called when Do is done.
Do(key string, fn func() (interface{}, error)) (interface{}, error)
}
// Stats are per-group statistics.
type Stats struct {
Gets AtomicInt // any Get request, including from peers
CacheHits AtomicInt // either cache was good
PeerLoads AtomicInt // either remote load or remote cache hit (not an error)
PeerErrors AtomicInt
Loads AtomicInt // (gets - cacheHits)
LoadsDeduped AtomicInt // after singleflight
LocalLoads AtomicInt // total good local loads
LocalLoadErrs AtomicInt // total bad local loads
ServerRequests AtomicInt // gets that came over the network from peers
}
// Name returns the name of the group.
func (g *Group) Name() string {
return g.name
}
func (g *Group) initPeers() {
if g.peers == nil {
g.peers = getPeers()
}
}
func (g *Group) Get(ctx Context, key string, dest Sink) error {
g.peersOnce.Do(g.initPeers)
g.Stats.Gets.Add(1)
if dest == nil {
return errors.New("groupcache: nil dest Sink")
}
value, cacheHit := g.lookupCache(key)
if cacheHit {
g.Stats.CacheHits.Add(1)
return setSinkView(dest, value)
}
// Optimization to avoid double unmarshalling or copying: keep
// track of whether the dest was already populated. One caller
// (if local) will set this; the losers will not. The common
// case will likely be one caller.
destPopulated := false
value, destPopulated, err := g.load(ctx, key, dest)
if err != nil {
return err
}
if destPopulated {
return nil
}
return setSinkView(dest, value)
}
// load loads key either by invoking the getter locally or by sending it to another machine.
func (g *Group) load(ctx Context, key string, dest Sink) (value ByteView, destPopulated bool, err error) {
g.Stats.Loads.Add(1)
viewi, err := g.loadGroup.Do(key, func() (interface{}, error) {
// Check the cache again because singleflight can only dedup calls
// that overlap concurrently. It's possible for 2 concurrent
// requests to miss the cache, resulting in 2 load() calls. An
// unfortunate goroutine scheduling would result in this callback
// being run twice, serially. If we don't check the cache again,
// cache.nbytes would be incremented below even though there will
// be only one entry for this key.
//
// Consider the following serialized event ordering for two
// goroutines in which this callback gets called twice for hte
// same key:
// 1: Get("key")
// 2: Get("key")
// 1: lookupCache("key")
// 2: lookupCache("key")
// 1: load("key")
// 2: load("key")
// 1: loadGroup.Do("key", fn)
// 1: fn()
// 2: loadGroup.Do("key", fn)
// 2: fn()
if value, cacheHit := g.lookupCache(key); cacheHit {
g.Stats.CacheHits.Add(1)
return value, nil
}
g.Stats.LoadsDeduped.Add(1)
var value ByteView
var err error
if peer, ok := g.peers.PickPeer(key); ok {
value, err = g.getFromPeer(ctx, peer, key)
if err == nil {
g.Stats.PeerLoads.Add(1)
return value, nil
}
g.Stats.PeerErrors.Add(1)
// TODO(bradfitz): log the peer's error? keep
// log of the past few for /groupcachez? It's
// probably boring (normal task movement), so not
// worth logging I imagine.
}
value, err = g.getLocally(ctx, key, dest)
if err != nil {
g.Stats.LocalLoadErrs.Add(1)
return nil, err
}
g.Stats.LocalLoads.Add(1)
destPopulated = true // only one caller of load gets this return value
g.populateCache(key, value, &g.mainCache)
return value, nil
})
if err == nil {
value = viewi.(ByteView)
}
return
}
func (g *Group) getLocally(ctx Context, key string, dest Sink) (ByteView, error) {
err := g.getter.Get(ctx, key, dest)
if err != nil {
return ByteView{}, err
}
return dest.view()
}
func (g *Group) getFromPeer(ctx Context, peer ProtoGetter, key string) (ByteView, error) {
req := &pb.GetRequest{
Group: &g.name,
Key: &key,
}
res := &pb.GetResponse{}
err := peer.Get(ctx, req, res)
if err != nil {
return ByteView{}, err
}
value := ByteView{b: res.Value}
// TODO(bradfitz): use res.MinuteQps or something smart to
// conditionally populate hotCache. For now just do it some
// percentage of the time.
if rand.Intn(10) == 0 {
g.populateCache(key, value, &g.hotCache)
}
return value, nil
}
func (g *Group) lookupCache(key string) (value ByteView, ok bool) {
if g.cacheBytes <= 0 {
return
}
value, ok = g.mainCache.get(key)
if ok {
return
}
value, ok = g.hotCache.get(key)
return
}
func (g *Group) populateCache(key string, value ByteView, cache *cache) {
if g.cacheBytes <= 0 {
return
}
cache.add(key, value)
// Evict items from cache(s) if necessary.
for {
mainBytes := g.mainCache.bytes()
hotBytes := g.hotCache.bytes()
if mainBytes+hotBytes <= g.cacheBytes {
return
}
// TODO(bradfitz): this is good-enough-for-now logic.
// It should be something based on measurements and/or
// respecting the costs of different resources.
victim := &g.mainCache
if hotBytes > mainBytes/8 {
victim = &g.hotCache
}
victim.removeOldest()
}
}
// CacheType represents a type of cache.
type CacheType int
const (
// The MainCache is the cache for items that this peer is the
// owner for.
MainCache CacheType = iota + 1
// The HotCache is the cache for items that seem popular
// enough to replicate to this node, even though it's not the
// owner.
HotCache
)
// CacheStats returns stats about the provided cache within the group.
func (g *Group) CacheStats(which CacheType) CacheStats {
switch which {
case MainCache:
return g.mainCache.stats()
case HotCache:
return g.hotCache.stats()
default:
return CacheStats{}
}
}
// cache is a wrapper around an *lru.Cache that adds synchronization,
// makes values always be ByteView, and counts the size of all keys and
// values.
type cache struct {
mu sync.RWMutex
nbytes int64 // of all keys and values
lru *lru.Cache
nhit, nget int64
nevict int64 // number of evictions
}
func (c *cache) stats() CacheStats {
c.mu.RLock()
defer c.mu.RUnlock()
return CacheStats{
Bytes: c.nbytes,
Items: c.itemsLocked(),
Gets: c.nget,
Hits: c.nhit,
Evictions: c.nevict,
}
}
func (c *cache) add(key string, value ByteView) {
c.mu.Lock()
defer c.mu.Unlock()
if c.lru == nil {
c.lru = &lru.Cache{
OnEvicted: func(key lru.Key, value interface{}) {
val := value.(ByteView)
c.nbytes -= int64(len(key.(string))) + int64(val.Len())
c.nevict++
},
}
}
c.lru.Add(key, value)
c.nbytes += int64(len(key)) + int64(value.Len())
}
func (c *cache) get(key string) (value ByteView, ok bool) {
c.mu.Lock()
defer c.mu.Unlock()
c.nget++
if c.lru == nil {
return
}
vi, ok := c.lru.Get(key)
if !ok {
return
}
c.nhit++
return vi.(ByteView), true
}
func (c *cache) removeOldest() {
c.mu.Lock()
defer c.mu.Unlock()
if c.lru != nil {
c.lru.RemoveOldest()
}
}
func (c *cache) bytes() int64 {
c.mu.RLock()
defer c.mu.RUnlock()
return c.nbytes
}
func (c *cache) items() int64 {
c.mu.RLock()
defer c.mu.RUnlock()
return c.itemsLocked()
}
func (c *cache) itemsLocked() int64 {
if c.lru == nil {
return 0
}
return int64(c.lru.Len())
}
// An AtomicInt is an int64 to be accessed atomically.
type AtomicInt int64
// Add atomically adds n to i.
func (i *AtomicInt) Add(n int64) {
atomic.AddInt64((*int64)(i), n)
}
// Get atomically gets the value of i.
func (i *AtomicInt) Get() int64 {
return atomic.LoadInt64((*int64)(i))
}
func (i *AtomicInt) String() string {
return strconv.FormatInt(i.Get(), 10)
}
// CacheStats are returned by stats accessors on Group.
type CacheStats struct {
Bytes int64
Items int64
Gets int64
Hits int64
Evictions int64
}