README
¶
groupcache
Summary
groupcache is a caching and cache-filling library, intended as a replacement for memcached in many cases.
For API docs and examples, see http://godoc.org/github.com/golang/groupcache
Comparison to memcached
Like memcached, groupcache:
- shards by key to select which peer is responsible for that key
Unlike memcached, groupcache:
-
does not require running a separate set of servers, thus massively reducing deployment/configuration pain. groupcache is a client library as well as a server. It connects to its own peers.
-
comes with a cache filling mechanism. Whereas memcached just says "Sorry, cache miss", often resulting in a thundering herd of database (or whatever) loads from an unbounded number of clients (which has resulted in several fun outages), groupcache coordinates cache fills such that only one load in one process of an entire replicated set of processes populates the cache, then multiplexes the loaded value to all callers.
-
does not support versioned values. If key "foo" is value "bar", key "foo" must always be "bar". There are neither cache expiration times, nor explicit cache evictions. Thus there is also no CAS, nor Increment/Decrement. This also means that groupcache....
-
... supports automatic mirroring of super-hot items to multiple processes. This prevents memcached hot spotting where a machine's CPU and/or NIC are overloaded by very popular keys/values.
-
is currently only available for Go. It's very unlikely that I (bradfitz@) will port the code to any other language.
Loading process
In a nutshell, a groupcache lookup of Get("foo") looks like:
(On machine #5 of a set of N machines running the same code)
-
Is the value of "foo" in local memory because it's super hot? If so, use it.
-
Is the value of "foo" in local memory because peer #5 (the current peer) is the owner of it? If so, use it.
-
Amongst all the peers in my set of N, am I the owner of the key "foo"? (e.g. does it consistent hash to 5?) If so, load it. If other callers come in, via the same process or via RPC requests from peers, they block waiting for the load to finish and get the same answer. If not, RPC to the peer that's the owner and get the answer. If the RPC fails, just load it locally (still with local dup suppression).
Users
groupcache is in production use by dl.google.com (its original user), parts of Blogger, parts of Google Code, parts of Google Fiber, parts of Google production monitoring systems, etc.
Presentations
See http://talks.golang.org/2013/oscon-dl.slide
Help
Use the golang-nuts mailing list for any discussion or questions.
Documentation
¶
Overview ¶
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.
Index ¶
- func RegisterNewGroupHook(fn func(*Group))
- func RegisterPeerPicker(fn func() PeerPicker)
- func RegisterServerStart(fn func())
- type AtomicInt
- type ByteView
- func (v ByteView) At(i int) byte
- func (v ByteView) ByteSlice() []byte
- func (v ByteView) Copy(dest []byte) int
- func (v ByteView) Equal(b2 ByteView) bool
- func (v ByteView) EqualBytes(b2 []byte) bool
- func (v ByteView) EqualString(s string) bool
- func (v ByteView) Len() int
- func (v ByteView) ReadAt(p []byte, off int64) (n int, err error)
- func (v ByteView) Reader() io.ReadSeeker
- func (v ByteView) Slice(from, to int) ByteView
- func (v ByteView) SliceFrom(from int) ByteView
- func (v ByteView) String() string
- type CacheStats
- type CacheType
- type Context
- type Getter
- type GetterFunc
- type Group
- type HTTPPool
- type NoPeers
- type PeerPicker
- type ProtoGetter
- type Sink
- type Stats
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func RegisterNewGroupHook ¶
func RegisterNewGroupHook(fn func(*Group))
RegisterNewGroupHook registers a hook that is run each time a group is created.
func RegisterPeerPicker ¶
func RegisterPeerPicker(fn func() PeerPicker)
RegisterPeerPicker registers the peer initialization function. It is called once, when the first group is created.
func RegisterServerStart ¶
func RegisterServerStart(fn func())
RegisterServerStart registers a hook that is run when the first group is created.
Types ¶
type ByteView ¶
type ByteView struct {
// contains filtered or unexported fields
}
A ByteView holds an immutable view of bytes. Internally it wraps either a []byte or a string, but that detail is invisible to callers.
A ByteView is meant to be used as a value type, not a pointer (like a time.Time).
func (ByteView) EqualBytes ¶
EqualBytes returns whether the bytes in b are the same as the bytes in b2.
func (ByteView) EqualString ¶
EqualString returns whether the bytes in b are the same as the bytes in s.
func (ByteView) Reader ¶
func (v ByteView) Reader() io.ReadSeeker
Reader returns an io.ReadSeeker for the bytes in v.
type CacheStats ¶
CacheStats are returned by stats accessors on Group.
type Context ¶
type Context interface{}
Context is an opaque value passed through calls to the ProtoGetter. It may be nil if your ProtoGetter implementation does not require a context.
type Getter ¶
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 Getter loads data for a key.
type GetterFunc ¶
A GetterFunc implements Getter with a function.
type Group ¶
type Group struct {
// Stats are statistics on the group.
Stats Stats
// contains filtered or unexported fields
}
A Group is a cache namespace and associated data loaded spread over a group of 1 or more machines.
func GetGroup ¶
GetGroup returns the named group previously created with NewGroup, or nil if there's no such group.
func NewGroup ¶
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 (*Group) CacheStats ¶
func (g *Group) CacheStats(which CacheType) CacheStats
CacheStats returns stats about the provided cache within the group.
type HTTPPool ¶
type HTTPPool struct {
// Context optionally specifies a context for the server to use when it
// receives a request.
// If nil, the server uses a nil Context.
Context func(*http.Request) Context
// Transport optionally specifies an http.RoundTripper for the client
// to use when it makes a request.
// If nil, the client uses http.DefaultTransport.
Transport func(Context) http.RoundTripper
// contains filtered or unexported fields
}
HTTPPool implements PeerPicker for a pool of HTTP peers.
func NewHTTPPool ¶
NewHTTPPool initializes an HTTP pool of peers. It registers itself as a PeerPicker and as an HTTP handler with the http.DefaultServeMux. The self argument be a valid base URL that points to the current server, for example "http://example.net:8000".
func (*HTTPPool) Set ¶
Set updates the pool's list of peers. Each peer value should be a valid base URL, for example "http://example.net:8000".
type NoPeers ¶
type NoPeers struct{}
NoPeers is an implementation of PeerPicker that never finds a peer.
type PeerPicker ¶
type PeerPicker interface {
// PickPeer returns the peer that owns the specific key
// and true to indicate that a remote peer was nominated.
// It returns nil, false if the key owner is the current peer.
PickPeer(key string) (peer ProtoGetter, ok bool)
}
PeerPicker is the interface that must be implemented to locate the peer that owns a specific key.
type ProtoGetter ¶
type ProtoGetter interface {
Get(context Context, in *pb.GetRequest, out *pb.GetResponse) error
}
ProtoGetter is the interface that must be implemented by a peer.
type Sink ¶
type Sink interface {
// SetString sets the value to s.
SetString(s string) error
// SetBytes sets the value to the contents of v.
// The caller retains ownership of v.
SetBytes(v []byte) error
// SetProto sets the value to the encoded version of m.
// The caller retains ownership of m.
SetProto(m proto.Message) error
// contains filtered or unexported methods
}
A Sink receives data from a Get call.
Implementation of Getter must call exactly one of the Set methods on success.
func AllocatingByteSliceSink ¶
AllocatingByteSliceSink returns a Sink that allocates a byte slice to hold the received value and assigns it to *dst. The memory is not retained by groupcache.
func ByteViewSink ¶
ByteViewSink returns a Sink that populates a ByteView.
func StringSink ¶
StringSink returns a Sink that populates the provided string pointer.
func TruncatingByteSliceSink ¶
TruncatingByteSliceSink returns a Sink that writes up to len(*dst) bytes to *dst. If more bytes are available, they're silently truncated. If fewer bytes are available than len(*dst), *dst is shrunk to fit the number of bytes available.
type Stats ¶
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
}
Stats are per-group statistics.
Source Files
Directories
¶
| Path | Synopsis |
|---|---|
|
Package consistenthash provides an implementation of a ring hash.
|
Package consistenthash provides an implementation of a ring hash. |
|
Package lru implements an LRU cache.
|
Package lru implements an LRU cache. |
|
Package singleflight provides a duplicate function call suppression mechanism.
|
Package singleflight provides a duplicate function call suppression mechanism. |