lease.go
package clientv3
import (
"sync"
"time"
"github.com/coreos/etcd/etcdserver/api/v3rpc/rpctypes"
pb "github.com/coreos/etcd/etcdserver/etcdserverpb"
"golang.org/x/net/context"
"google.golang.org/grpc"
)
type (
LeaseRevokeResponse pb.LeaseRevokeResponse
LeaseID int64
)
// LeaseGrantResponse is used to convert the protobuf grant response.
type LeaseGrantResponse struct {
*pb.ResponseHeader
ID LeaseID
TTL int64
Error string
}
// LeaseKeepAliveResponse is used to convert the protobuf keepalive response.
type LeaseKeepAliveResponse struct {
*pb.ResponseHeader
ID LeaseID
TTL int64
}
// LeaseTimeToLiveResponse is used to convert the protobuf lease timetolive response.
type LeaseTimeToLiveResponse struct {
*pb.ResponseHeader
ID LeaseID `json:"id"`
// TTL is the remaining TTL in seconds for the lease; the lease will expire in under TTL+1 seconds.
TTL int64 `json:"ttl"`
// GrantedTTL is the initial granted time in seconds upon lease creation/renewal.
GrantedTTL int64 `json:"granted-ttl"`
// Keys is the list of keys attached to this lease.
Keys [][]byte `json:"keys"`
}
const (
// defaultTTL is the assumed lease TTL used for the first keepalive
// deadline before the actual TTL is known to the client.
defaultTTL = 5 * time.Second
// a small buffer to store unsent lease responses.
leaseResponseChSize = 16
// NoLease is a lease ID for the absence of a lease.
NoLease LeaseID = 0
)
type Lease interface {
// Grant creates a new lease.
Grant(ctx context.Context, ttl int64) (*LeaseGrantResponse, error)
// Revoke revokes the given lease.
Revoke(ctx context.Context, id LeaseID) (*LeaseRevokeResponse, error)
// TimeToLive retrieves the lease information of the given lease ID.
TimeToLive(ctx context.Context, id LeaseID, opts ...LeaseOption) (*LeaseTimeToLiveResponse, error)
// KeepAlive keeps the given lease alive forever.
KeepAlive(ctx context.Context, id LeaseID) (<-chan *LeaseKeepAliveResponse, error)
// KeepAliveOnce renews the lease once. In most of the cases, Keepalive
// should be used instead of KeepAliveOnce.
KeepAliveOnce(ctx context.Context, id LeaseID) (*LeaseKeepAliveResponse, error)
// Close releases all resources Lease keeps for efficient communication
// with the etcd server.
Close() error
}
type lessor struct {
mu sync.Mutex // guards all fields
// donec is closed when recvKeepAliveLoop stops
donec chan struct{}
remote pb.LeaseClient
stream pb.Lease_LeaseKeepAliveClient
streamCancel context.CancelFunc
stopCtx context.Context
stopCancel context.CancelFunc
keepAlives map[LeaseID]*keepAlive
// firstKeepAliveTimeout is the timeout for the first keepalive request
// before the actual TTL is known to the lease client
firstKeepAliveTimeout time.Duration
}
// keepAlive multiplexes a keepalive for a lease over multiple channels
type keepAlive struct {
chs []chan<- *LeaseKeepAliveResponse
ctxs []context.Context
// deadline is the time the keep alive channels close if no response
deadline time.Time
// nextKeepAlive is when to send the next keep alive message
nextKeepAlive time.Time
// donec is closed on lease revoke, expiration, or cancel.
donec chan struct{}
}
func NewLease(c *Client) Lease {
l := &lessor{
donec: make(chan struct{}),
keepAlives: make(map[LeaseID]*keepAlive),
remote: RetryLeaseClient(c),
firstKeepAliveTimeout: c.cfg.DialTimeout + time.Second,
}
if l.firstKeepAliveTimeout == time.Second {
l.firstKeepAliveTimeout = defaultTTL
}
l.stopCtx, l.stopCancel = context.WithCancel(context.Background())
go l.recvKeepAliveLoop()
go l.deadlineLoop()
return l
}
func (l *lessor) Grant(ctx context.Context, ttl int64) (*LeaseGrantResponse, error) {
cctx, cancel := context.WithCancel(ctx)
done := cancelWhenStop(cancel, l.stopCtx.Done())
defer close(done)
for {
r := &pb.LeaseGrantRequest{TTL: ttl}
resp, err := l.remote.LeaseGrant(cctx, r)
if err == nil {
gresp := &LeaseGrantResponse{
ResponseHeader: resp.GetHeader(),
ID: LeaseID(resp.ID),
TTL: resp.TTL,
Error: resp.Error,
}
return gresp, nil
}
if isHaltErr(cctx, err) {
return nil, toErr(cctx, err)
}
if nerr := l.newStream(); nerr != nil {
return nil, nerr
}
}
}
func (l *lessor) Revoke(ctx context.Context, id LeaseID) (*LeaseRevokeResponse, error) {
cctx, cancel := context.WithCancel(ctx)
done := cancelWhenStop(cancel, l.stopCtx.Done())
defer close(done)
for {
r := &pb.LeaseRevokeRequest{ID: int64(id)}
resp, err := l.remote.LeaseRevoke(cctx, r)
if err == nil {
return (*LeaseRevokeResponse)(resp), nil
}
if isHaltErr(ctx, err) {
return nil, toErr(ctx, err)
}
if nerr := l.newStream(); nerr != nil {
return nil, nerr
}
}
}
func (l *lessor) TimeToLive(ctx context.Context, id LeaseID, opts ...LeaseOption) (*LeaseTimeToLiveResponse, error) {
cctx, cancel := context.WithCancel(ctx)
done := cancelWhenStop(cancel, l.stopCtx.Done())
defer close(done)
for {
r := toLeaseTimeToLiveRequest(id, opts...)
resp, err := l.remote.LeaseTimeToLive(cctx, r, grpc.FailFast(false))
if err == nil {
gresp := &LeaseTimeToLiveResponse{
ResponseHeader: resp.GetHeader(),
ID: LeaseID(resp.ID),
TTL: resp.TTL,
GrantedTTL: resp.GrantedTTL,
Keys: resp.Keys,
}
return gresp, nil
}
if isHaltErr(cctx, err) {
return nil, toErr(cctx, err)
}
}
}
func (l *lessor) KeepAlive(ctx context.Context, id LeaseID) (<-chan *LeaseKeepAliveResponse, error) {
ch := make(chan *LeaseKeepAliveResponse, leaseResponseChSize)
l.mu.Lock()
ka, ok := l.keepAlives[id]
if !ok {
// create fresh keep alive
ka = &keepAlive{
chs: []chan<- *LeaseKeepAliveResponse{ch},
ctxs: []context.Context{ctx},
deadline: time.Now().Add(l.firstKeepAliveTimeout),
nextKeepAlive: time.Now(),
donec: make(chan struct{}),
}
l.keepAlives[id] = ka
} else {
// add channel and context to existing keep alive
ka.ctxs = append(ka.ctxs, ctx)
ka.chs = append(ka.chs, ch)
}
l.mu.Unlock()
go l.keepAliveCtxCloser(id, ctx, ka.donec)
return ch, nil
}
func (l *lessor) KeepAliveOnce(ctx context.Context, id LeaseID) (*LeaseKeepAliveResponse, error) {
cctx, cancel := context.WithCancel(ctx)
done := cancelWhenStop(cancel, l.stopCtx.Done())
defer close(done)
for {
resp, err := l.keepAliveOnce(cctx, id)
if err == nil {
if resp.TTL == 0 {
err = rpctypes.ErrLeaseNotFound
}
return resp, err
}
if isHaltErr(ctx, err) {
return nil, toErr(ctx, err)
}
if nerr := l.newStream(); nerr != nil {
return nil, nerr
}
}
}
func (l *lessor) Close() error {
l.stopCancel()
<-l.donec
return nil
}
func (l *lessor) keepAliveCtxCloser(id LeaseID, ctx context.Context, donec <-chan struct{}) {
select {
case <-donec:
return
case <-l.donec:
return
case <-ctx.Done():
}
l.mu.Lock()
defer l.mu.Unlock()
ka, ok := l.keepAlives[id]
if !ok {
return
}
// close channel and remove context if still associated with keep alive
for i, c := range ka.ctxs {
if c == ctx {
close(ka.chs[i])
ka.ctxs = append(ka.ctxs[:i], ka.ctxs[i+1:]...)
ka.chs = append(ka.chs[:i], ka.chs[i+1:]...)
break
}
}
// remove if no one more listeners
if len(ka.chs) == 0 {
delete(l.keepAlives, id)
}
}
func (l *lessor) keepAliveOnce(ctx context.Context, id LeaseID) (*LeaseKeepAliveResponse, error) {
cctx, cancel := context.WithCancel(ctx)
defer cancel()
stream, err := l.remote.LeaseKeepAlive(cctx, grpc.FailFast(false))
if err != nil {
return nil, toErr(ctx, err)
}
err = stream.Send(&pb.LeaseKeepAliveRequest{ID: int64(id)})
if err != nil {
return nil, toErr(ctx, err)
}
resp, rerr := stream.Recv()
if rerr != nil {
return nil, toErr(ctx, rerr)
}
karesp := &LeaseKeepAliveResponse{
ResponseHeader: resp.GetHeader(),
ID: LeaseID(resp.ID),
TTL: resp.TTL,
}
return karesp, nil
}
func (l *lessor) recvKeepAliveLoop() {
defer func() {
l.mu.Lock()
close(l.donec)
for _, ka := range l.keepAlives {
ka.Close()
}
l.keepAlives = make(map[LeaseID]*keepAlive)
l.mu.Unlock()
}()
stream, serr := l.resetRecv()
for serr == nil {
resp, err := stream.Recv()
if err != nil {
if isHaltErr(l.stopCtx, err) {
return
}
stream, serr = l.resetRecv()
continue
}
l.recvKeepAlive(resp)
}
}
// resetRecv opens a new lease stream and starts sending LeaseKeepAliveRequests
func (l *lessor) resetRecv() (pb.Lease_LeaseKeepAliveClient, error) {
if err := l.newStream(); err != nil {
return nil, err
}
stream := l.getKeepAliveStream()
go l.sendKeepAliveLoop(stream)
return stream, nil
}
// recvKeepAlive updates a lease based on its LeaseKeepAliveResponse
func (l *lessor) recvKeepAlive(resp *pb.LeaseKeepAliveResponse) {
karesp := &LeaseKeepAliveResponse{
ResponseHeader: resp.GetHeader(),
ID: LeaseID(resp.ID),
TTL: resp.TTL,
}
l.mu.Lock()
defer l.mu.Unlock()
ka, ok := l.keepAlives[karesp.ID]
if !ok {
return
}
if karesp.TTL <= 0 {
// lease expired; close all keep alive channels
delete(l.keepAlives, karesp.ID)
ka.Close()
return
}
// send update to all channels
nextKeepAlive := time.Now().Add(1 + time.Duration(karesp.TTL/3)*time.Second)
ka.deadline = time.Now().Add(time.Duration(karesp.TTL) * time.Second)
for _, ch := range ka.chs {
select {
case ch <- karesp:
ka.nextKeepAlive = nextKeepAlive
default:
}
}
}
// deadlineLoop reaps any keep alive channels that have not received a response
// within the lease TTL
func (l *lessor) deadlineLoop() {
for {
select {
case <-time.After(time.Second):
case <-l.donec:
return
}
now := time.Now()
l.mu.Lock()
for id, ka := range l.keepAlives {
if ka.deadline.Before(now) {
// waited too long for response; lease may be expired
ka.Close()
delete(l.keepAlives, id)
}
}
l.mu.Unlock()
}
}
// sendKeepAliveLoop sends LeaseKeepAliveRequests for the lifetime of a lease stream
func (l *lessor) sendKeepAliveLoop(stream pb.Lease_LeaseKeepAliveClient) {
for {
select {
case <-time.After(500 * time.Millisecond):
case <-stream.Context().Done():
return
case <-l.donec:
return
case <-l.stopCtx.Done():
return
}
var tosend []LeaseID
now := time.Now()
l.mu.Lock()
for id, ka := range l.keepAlives {
if ka.nextKeepAlive.Before(now) {
tosend = append(tosend, id)
}
}
l.mu.Unlock()
for _, id := range tosend {
r := &pb.LeaseKeepAliveRequest{ID: int64(id)}
if err := stream.Send(r); err != nil {
// TODO do something with this error?
return
}
}
}
}
func (l *lessor) getKeepAliveStream() pb.Lease_LeaseKeepAliveClient {
l.mu.Lock()
defer l.mu.Unlock()
return l.stream
}
func (l *lessor) newStream() error {
sctx, cancel := context.WithCancel(l.stopCtx)
stream, err := l.remote.LeaseKeepAlive(sctx, grpc.FailFast(false))
if err != nil {
cancel()
return toErr(sctx, err)
}
l.mu.Lock()
defer l.mu.Unlock()
if l.stream != nil && l.streamCancel != nil {
l.stream.CloseSend()
l.streamCancel()
}
l.streamCancel = cancel
l.stream = stream
return nil
}
func (ka *keepAlive) Close() {
close(ka.donec)
for _, ch := range ka.chs {
close(ch)
}
}
// cancelWhenStop calls cancel when the given stopc fires. It returns a done chan. done
// should be closed when the work is finished. When done fires, cancelWhenStop will release
// its internal resource.
func cancelWhenStop(cancel context.CancelFunc, stopc <-chan struct{}) chan<- struct{} {
done := make(chan struct{}, 1)
go func() {
select {
case <-stopc:
case <-done:
}
cancel()
}()
return done
}
lease.go的更多相关文章
- 基于Lease分布式系统重试服务选举
/** * Copyright (c) 2015, www.cubbery.com. All rights reserved. */ package com.cubbery.event.retry; ...
- 分布式入门之1：Lease机制
引子: 分布式系统中,如何确认一个节点是否工作正常? 如果有3副本A.B.C,并通过中心结点M来管理.其中A为主副本. 未接触过分布式的直观的处理方法是在每个副本与中心节点M中维护一个心跳,期 ...
- Azure 删除VHD时报错：There is currently a lease on the blob and no lease ID was specified in the request
可下载:http://clumsyleaf.com/products/cloudxplorer 然后在Accounts中新建一个Account,账号与Key,可在相应的storage Manage A ...
- sudo -u hdfs hdfs balancer出现异常 No lease on /system/balancer.id
16/06/02 20:34:05 INFO balancer.Balancer: namenodes = [hdfs://dlhtHadoop101:8022, hdfs://dlhtHadoop1 ...
- Hey，man，are you ok? －－ 关于心跳、故障监测、lease机制
电话之于短信.微信的一个很大的不同点在于,前者更加及时,有更快速直接的反馈:而后面两个虽然称之为instant message,但经常时发出去了就得等对方回复,等多久是不确定的.打电话能明确知道对方在 ...
- Lease问题
经过查明原来是lease引发的问题.不过查问题的过程让我们耽误了很多修复故障的时间,很是不爽. 起因:datanode和regionserver以及master同时挂掉 现象:datanode重启后, ...
- 分析dhcp.lease文件，统计DHCP服务器IP自动分配
#!/usr/bin/env python # coding=utf-8 import string import time,datetime class TIMEFORMAT: def __init ...
- 深入NAS协议系列: 召唤SMB2 OpLock/Lease
这是从事存储行业十年以来我写的第一篇博客,希望借此开始把自己这些年所积累的一些干货借这个平台做分享. 虽然NAS协议众多,但核心的就那个几个:NFS,SMB/CIFS, FTP/SFTP, 其中SMB ...
- HDFS Lease Recovey 和 Block Recovery
这篇分析一下Lease Recovery 和 Block Recovery hdfs支持hflush后,需要保证hflush的数据被读到,datanode重启不能简单的丢弃文件的最后一个block,而 ...
随机推荐
- MySQL 表名区分大小写设置
1.关闭MySQL服务: 控制面板主页-管理工具-服务-MySQL服务 2.在服务器运行目录找到my.ini 或者my.cnf文件: 在[mysqld]下面增加一行添加 :lower_ ...
- html5中新增的元素和废除的元素
一.新增的结构元素 1.section元素表示页面中的一个内容区块,比如章节.页眉.页脚或页面中的其他部分.它可以与h1.h2.h3.h4.h5.h6等元素结合起来使用,标示文档结构. h5中的代码事 ...
- 《转》xcode创建一个工程的多个taget,便于测试和发布多个版本
背景:很多时候,我们需要在一个工程中创立多个target,也就是说我们希望同一份代码可以创建两个应用,放到模拟器或者真机上,或者是,我们平时有N多人合作开发,当测试的时候,在A这里装了一遍测A写的那块 ...
- 修改flume源码，使其HTTPSource具备访问路径功能
目前有一个需求,就是Flume可以作为一个类似于tomcat的服务器,可以通过post请求进行访问,并且路径需要:ip:port/contextPath格式. 经过一些资料获悉,httpSource只 ...
- WEB 集群与负载均衡(一)基本概念-上
Web集群是由多个同时运行同一个web应用的服务器组成,在外界看来就像一个服务器一样,这多台服务器共同来为客户提供更高性能的服务.集群更标准的定义是:一组相互独立的服务器在网络中表现为单一的系统,并以 ...
- 自动生成MyEclipse 安装破解码
新建一个class 文件,Debug 模式运行一个,输入任意值 ,回车得到破解安装码 代码文件如下: import java.io.*; public class MyEclipseGen { pri ...
- 原生Eclipse下Java服务器调试的一个问题
当你对Server的配置修改以后,最好到 workspacedir\.metadata\.plugins\org.eclipse.wst.server.core\tmp0目录下把缓存文件给删除了,否则 ...
- Eclipse+Resin开发环境迁移中发生的一些问题
换新机器了,系统也从XP升级到64位WIn7.某些旧工具直接无法用了.下面简单谈一下标题的内容 1.非泛型的容器类引入在JDK1.7以下编译好像已经不行了.比如Java.util.ArrayList这 ...
- Android设计开发笔记
1.因为Android的开发是基于框架的开发:往对方指定的位置加代码:其运行的Message\Handler机制也决定了其单步跟踪也不方便,所以建立新代码时要多Log,这样不但便于调试,而且帮助你加深 ...
- 关于Python的super用法研究
一.问题的发现与提出 在Python类的方法(method)中,要调用父类的某个方法,在python 2.2以前,通常的写法如代码段1: 代码段1: class A: def __init__(se ...