第一步,使用git到如下地址拉取项目
$ git clone git://g.csail.mit.edu/6.824-golabs-2020 6.824
$ cd 6.824
$ ls
Makefile src
$
执行如下命令
$ cd ~/6.824
$ cd src/main
$ go build -buildmode=plugin ../mrapps/wc.go
$ rm mr-out*
$ go run mrsequential.go wc.so pg*.txt
$ more mr-out-0
A 509
ABOUT 2
ACT 8
...
发现报错,是go mod的问题, 执行以下操作修改go mod(去stakoverflow寻找解决方案)
cd 6.824
go mod init "6.824-golabs-2020"
# change file src/mrapps/wc.go line9 to `import "6.824-golabs-2020/src/mr"`
cd src
go build -buildmode=plugin mrapps/wc.go
还需要修改mrsequential.go 的导入包,最后测试成功,完成环境搭建。
下面来看mrsequential.go的主函数
第一个参数wc.so提供了map函数和reduce函数,第二个之后的参数是一系列的文件名
func main() {
if len(os.Args) < 3 {
fmt.Fprintf(os.Stderr, "Usage: mrsequential xxx.so inputfiles...\n")
os.Exit(1)
}
mapf, reducef := loadPlugin(os.Args[1])
//
// read each input file,
// pass it to Map,
// accumulate the intermediate Map output.
//
intermediate := []mr.KeyValue{}
for _, filename := range os.Args[2:] {
file, err := os.Open(filename)
if err != nil {
log.Fatalf("cannot open %v", filename)
}
content, err := ioutil.ReadAll(file)
if err != nil {
log.Fatalf("cannot read %v", filename)
}
file.Close()
kva := mapf(filename, string(content))
intermediate = append(intermediate, kva...)
}
//
// a big difference from real MapReduce is that all the
// intermediate data is in one place, intermediate[],
// rather than being partitioned into NxM buckets.
//
sort.Sort(ByKey(intermediate))
oname := "mr-out-0"
ofile, _ := os.Create(oname)
//
// call Reduce on each distinct key in intermediate[],
// and print the result to mr-out-0.
//
i := 0
for i < len(intermediate) {
j := i + 1
for j < len(intermediate) && intermediate[j].Key == intermediate[i].Key {
j++
}
values := []string{}
for k := i; k < j; k++ {
values = append(values, intermediate[k].Value)
}
output := reducef(intermediate[i].Key, values)
// this is the correct format for each line of Reduce output.
fmt.Fprintf(ofile, "%v %v\n", intermediate[i].Key, output)
i = j
}
ofile.Close()
}
再去看一下wc.go
//
// The map function is called once for each file of input. The first
// argument is the name of the input file, and the second is the
// file's complete contents. You should ignore the input file name,
// and look only at the contents argument. The return value is a slice
// of key/value pairs.
//
func Map(filename string, contents string) []mr.KeyValue {
// function to detect word separators.
ff := func(r rune) bool { return !unicode.IsLetter(r) }
// split contents into an array of words.
words := strings.FieldsFunc(contents, ff)
kva := []mr.KeyValue{}
for _, w := range words {
kv := mr.KeyValue{w, "1"}
kva = append(kva, kv)
}
return kva
}
//
// The reduce function is called once for each key generated by the
// map tasks, with a list of all the values created for that key by
// any map task.
//
func Reduce(key string, values []string) string {
// return the number of occurrences of this word.
return strconv.Itoa(len(values))
}
mrmaster.go 主函数
func main() {
if len(os.Args) < 2 {
fmt.Fprintf(os.Stderr, "Usage: mrmaster inputfiles...\n")
os.Exit(1)
}
m := mr.MakeMaster(os.Args[1:], 10)
for m.Done() == false {
time.Sleep(time.Second)
}
time.Sleep(time.Second)
}
mrwork.go 主函数
func main() {
if len(os.Args) != 2 {
fmt.Fprintf(os.Stderr, "Usage: mrworker xxx.so\n")
os.Exit(1)
}
mapf, reducef := loadPlugin(os.Args[1])
mr.Worker(mapf, reducef)
}
首先做这一步
- One way to get started is to modify mr/worker.go's Worker() to send an RPC to the master asking for a task. Then modify the master to respond with the file name of an as-yet-unstarted map task. Then modify the worker to read that file and call the application Map function, as in mrsequential.go.
先测试master和worker的RPC
go run mrmaster.go pg-*.txt
go run mrworker.go wc.so
在这个RPC的基础上,实现worker的注册,存在好多个worker,所以worker必须有一个编号,和它对应的map,reduce函数(后面实现)
type worker struct {
workerID int
mapf func(string, string) []KeyValue
reducef func(string, []string) string
}
实现远程过程调用
func (w worker) callRegisterWorker() {
args := &RegisterArgs{}
reply := &RegisterReply{}
if ok := call("Master.RegisterWorker", args, reply); !ok {
log.Fatal("error: register worker failed.")
}
w.workerID = reply.ID
fmt.Printf("worker %v registered success.\n", w.workerID)
}
由于master要给worker分配编号,所以master要记录当前worker数,考虑到并发问题,还需要加锁
type Master struct {
// Your definitions here.
numWorkers int
mutex sync.Mutex
}
func (m *Master) RegisterWorker(args *RegisterArgs, reply *RegisterReply) error {
m.mutex.Lock()
defer m.mutex.Unlock()
reply.ID = m.numWorkers
m.numWorkers++
return nil
}
运行测试,未完待续。