Go Implements Mass Log Collection System (3)

I've reorganized the box for this log collection system, as shown below

The overall logic of the code to be implemented this time is:

The full code address is: https://github.com/pythonsite/logagent

Introduction to etcd

Highly available distributed key-value storage for configuring sharing and service discovery

Similar projects: zookeeper and consul

Development Language: go

Interface: Provides a restful interface for easy use

Implementation algorithm: Strong consistency based on raft algorithm, highly available service store directory

Scenarios for etcd:

Service Discovery and Service Registration
Configuration Center (needed by the log collection client we implemented)
Distributed Lock
master election

The official website has a very concise introduction to etcd:

etcd build:
Download address: https://github.com/coreos/etcd/releases/
Load the corresponding version in your environment and start it up

After startup, you can verify it with the following command:

[root@localhost etcd-v3.2.18-linux-amd64]# ./etcdctl set name zhaofan 

zhaofan
[root@localhost etcd-v3.2.18-linux-amd64]# ./etcdctl get name
zhaofan
[root@localhost etcd-v3.2.18-linux-amd64]#

Introduction and use of context

In fact, this thing is translated as context management, so the role of context is to do what, there are two main roles:

Control goroutine timeout
Save Context Data

Understand this with a simple example:

package main

import (
    "fmt"
    "time"
    "net/http"
    "context"
    "io/ioutil"
)


type Result struct{
    r *http.Response
    err error
}

func process(){
    ctx,cancel := context.WithTimeout(context.Background(),2*time.Second)
    defer cancel()
    tr := &http.Transport{}
    client := &http.Client{Transport:tr}
    c := make(chan Result,1)
    req,err := http.NewRequest("GET","http://www.google.com",nil)
    if err != nil{
        fmt.Println("http request failed,err:",err)
        return
    }
    // If the request succeeds, the data is stored in the pipeline
    go func(){
        resp,err := client.Do(req)
        pack := Result{resp,err}
        c <- pack
    }()

    select{
    case <- ctx.Done():
        tr.CancelRequest(req)
        fmt.Println("timeout!")
    case res := <-c:
        defer res.r.Body.Close()
        out,_:= ioutil.ReadAll(res.r.Body)
        fmt.Printf("server response:%s",out)
    }
    return

}

func main() {
    process()
}

Write a code example that saves context through context, such as:

package main

import (
    "github.com/Go-zh/net/context"
    "fmt"
)

func add(ctx context.Context,a,b int) int {
    traceId := ctx.Value("trace_id").(string)
    fmt.Printf("trace_id:%v\n",traceId)
    return a+b
}

func calc(ctx context.Context,a, b int) int{
    traceId := ctx.Value("trace_id").(string)
    fmt.Printf("trace_id:%v\n",traceId)
    //And will ctx Incoming to add in
    return add(ctx,a,b)
}

func main() {
    //take ctx Pass to calc in
    ctx := context.WithValue(context.Background(),"trace_id","123456")
    calc(ctx,20,30)

}

Use with etcd and context

For a simple example of connecting etcd via go: (Here is a small problem to note how etcd is started. The default startup may not connect, especially if you are installing it on a virtual basis, so you need to start it with the following commands:
./etcd --listen-client-urls http://0.0.0.0:2371 --advertise-client-urls http://0.0.0.0:2371 --listen-peer-urls http://0.0.0.0:2381
)

package main

import (
    etcd_client "github.com/coreos/etcd/clientv3"
    "time"
    "fmt"
)

func main() {
    cli, err := etcd_client.New(etcd_client.Config{
        Endpoints:[]string{"192.168.0.118:2371"},
        DialTimeout:5*time.Second,
    })
    if err != nil{
        fmt.Println("connect failed,err:",err)
        return
    }

    fmt.Println("connect success")
    defer cli.Close()
}

The following example is by connecting etcd s, storing values and taking values

package main

import (
    "github.com/coreos/etcd/clientv3"
    "time"
    "fmt"
    "context"
)

func main() {
    cli,err := clientv3.New(clientv3.Config{
        Endpoints:[]string{"192.168.0.118:2371"},
        DialTimeout:5*time.Second,
    })
    if err != nil{
        fmt.Println("connect failed,err:",err)
        return
    }
    fmt.Println("connect succ")
    defer cli.Close()
    ctx,cancel := context.WithTimeout(context.Background(),time.Second)
    _,err = cli.Put(ctx,"logagent/conf/","sample_value")
    cancel()
    if err != nil{
        fmt.Println("put failed,err",err)
        return
    }
    ctx, cancel = context.WithTimeout(context.Background(),time.Second)
    resp,err := cli.Get(ctx,"logagent/conf/")
    cancel()
    if err != nil{
        fmt.Println("get failed,err:",err)
        return
    }
    for _,ev := range resp.Kvs{
        fmt.Printf("%s:%s\n",ev.Key,ev.Value)
    }
}

There is also a useful example of context s for controlling the exit of an open goroutine with the following code:

package main

import (
    "context"
    "fmt"
)

func main() {
    // gen generates integers in a separate goroutine and
    // sends them to the returned channel.
    // The callers of gen need to cancel the context once
    // they are done consuming generated integers not to leak
    // the internal goroutine started by gen.
    gen := func(ctx context.Context) <-chan int {
        dst := make(chan int)
        n := 1
        go func() {
            for {
                select {
                case <-ctx.Done():
                    return // returning not to leak the goroutine
                case dst <- n:
                    n++
                }
            }
        }()
        return dst
    }

    ctx, cancel := context.WithCancel(context.Background())
    defer cancel() // cancel when we are finished consuming integers

    for n := range gen(ctx) {
        fmt.Println(n)
        if n == 5 {
            break
        }
    }
}

Code examples for WithDeadline demonstrations in official website documents:

package main


import (
    "context"
    "fmt"
    "time"
)

func main() {
    d := time.Now().Add(50 * time.Millisecond)
    ctx, cancel := context.WithDeadline(context.Background(), d)

    // Even though ctx will be expired, it is good practice to call its
    // cancelation function in any case. Failure to do so may keep the
    // context and its parent alive longer than necessary.
    defer cancel()

    select {
    case <-time.After(1 * time.Second):
        fmt.Println("overslept")
    case <-ctx.Done():
        fmt.Println(ctx.Err())
    }

}

With a basic use of the above code, if we use etcd for configuration management, how do we notify the corresponding server configuration changes after the configuration changes, as demonstrated by the following example:

package main

import (
    "github.com/coreos/etcd/clientv3"
    "time"
    "fmt"
    "context"
)

func main() {
    cli,err := clientv3.New(clientv3.Config{
        Endpoints:[]string{"192.168.0.118:2371"},
        DialTimeout:5*time.Second,
    })
    if err != nil {
        fmt.Println("connect failed,err:",err)
        return
    }
    defer cli.Close()
    // Blocked here
    rch := cli.Watch(context.Background(),"logagent/conf/")
    for wresp := range rch{
        for _,ev := range wresp.Events{
            fmt.Printf("%s %q : %q\n", ev.Type, ev.Kv.Key, ev.Kv.Value)
        }
    }
}

A simple example of implementing consumer code for kafka:

package main

import (
    "github.com/Shopify/sarama"
    "strings"
    "fmt"
    "time"
)

func main() {
    consumer,err := sarama.NewConsumer(strings.Split("192.168.0.118:9092",","),nil)
    if err != nil{
        fmt.Println("failed to start consumer:",err)
        return
    }
    partitionList,err := consumer.Partitions("nginx_log")
    if err != nil {
        fmt.Println("Failed to get the list of partitions:",err)
        return
    }
    fmt.Println(partitionList)
    for partition := range partitionList{
        pc,err := consumer.ConsumePartition("nginx_log",int32(partition),sarama.OffsetNewest)
        if err != nil {
            fmt.Printf("failed to start consumer for partition %d:%s\n",partition,err)
            return
        }
        defer pc.AsyncClose()
        go func(partitionConsumer sarama.PartitionConsumer){
            for msg := range pc.Messages(){
                fmt.Printf("partition:%d Offset:%d Key:%s Value:%s",msg.Partition,msg.Offset,string(msg.Key),string(msg.Value))
            }
        }(pc)
    }
    time.Sleep(time.Hour)
    consumer.Close()

}

But the above code is not the best code, because we finally waited for the goroutine to execute through time.sleep and we can change it to sync.WaitGroup

package main

import (
    "github.com/Shopify/sarama"
    "strings"
    "fmt"
    "sync"
)

var (
    wg sync.WaitGroup
)

func main() {
    consumer,err := sarama.NewConsumer(strings.Split("192.168.0.118:9092",","),nil)
    if err != nil{
        fmt.Println("failed to start consumer:",err)
        return
    }
    partitionList,err := consumer.Partitions("nginx_log")
    if err != nil {
        fmt.Println("Failed to get the list of partitions:",err)
        return
    }
    fmt.Println(partitionList)
    for partition := range partitionList{
        pc,err := consumer.ConsumePartition("nginx_log",int32(partition),sarama.OffsetNewest)
        if err != nil {
            fmt.Printf("failed to start consumer for partition %d:%s\n",partition,err)
            return
        }
        defer pc.AsyncClose()
        go func(partitionConsumer sarama.PartitionConsumer){
            wg.Add(1)
            for msg := range partitionConsumer.Messages(){
                fmt.Printf("partition:%d Offset:%d Key:%s Value:%s",msg.Partition,msg.Offset,string(msg.Key),string(msg.Value))
            }
            wg.Done()
        }(pc)
    }

    //time.Sleep(time.Hour)
    wg.Wait()
    consumer.Close()

}

Put the log information the client needs to collect into the etcd

The code for etcd processing is:

package main

import (
    "github.com/coreos/etcd/clientv3"
    "time"
    "github.com/astaxie/beego/logs"
    "context"
    "fmt"
)

var Client *clientv3.Client
var logConfChan chan string


// Initialization etcd
func initEtcd(addr []string,keyfmt string,timeout time.Duration)(err error){

    var keys []string
    for _,ip := range ipArrays{
        //keyfmt = /logagent/%s/log_config
        keys = append(keys,fmt.Sprintf(keyfmt,ip))
    }

    logConfChan = make(chan string,10)
    logs.Debug("etcd watch key:%v timeout:%v", keys, timeout)

    Client,err = clientv3.New(clientv3.Config{
        Endpoints:addr,
        DialTimeout: timeout,
    })
    if err != nil{
        logs.Error("connect failed,err:%v",err)
        return
    }
    logs.Debug("init etcd success")
    waitGroup.Add(1)
    for _, key := range keys{
        ctx,cancel := context.WithTimeout(context.Background(),2*time.Second)
        // from etcd Get information to collect logs in
        resp,err := Client.Get(ctx,key)
        cancel()
        if err != nil {
            logs.Warn("get key %s failed,err:%v",key,err)
            continue
        }

        for _, ev := range resp.Kvs{
            logs.Debug("%q : %q\n",  ev.Key, ev.Value)
            logConfChan <- string(ev.Value)
        }
    }
    go WatchEtcd(keys)
    return
}

func WatchEtcd(keys []string){
    // This is used to detect timely updates when log information needs to be collected changes
    var watchChans []clientv3.WatchChan
    for _,key := range keys{
        rch := Client.Watch(context.Background(),key)
        watchChans = append(watchChans,rch)
    }

    for {
        for _,watchC := range watchChans{
            select{
            case wresp := <-watchC:
                for _,ev:= range wresp.Events{
                    logs.Debug("%s %q : %q\n", ev.Type, ev.Kv.Key, ev.Kv.Value)
                    logConfChan <- string(ev.Kv.Value)
                }
            default:

            }
        }
        time.Sleep(time.Second)
    }
    waitGroup.Done()
}

func GetLogConf()chan string{
    return logConfChan
}

Likewise, speed-limiting processing has been added here, since the log collector cannot affect the performance of the current business, so limit.go has been added to limit speed:

package main

import (
    "time"
    "sync/atomic"
    "github.com/astaxie/beego/logs"
)

type SecondLimit struct {
    unixSecond int64
    curCount int32
    limit int32
}

func NewSecondLimit(limit int32) *SecondLimit {
    secLimit := &SecondLimit{
        unixSecond:time.Now().Unix(),
        curCount:0,
        limit:limit,
    }
    return secLimit
}

func (s *SecondLimit) Add(count int) {
    sec := time.Now().Unix()
    if sec == s.unixSecond {
        atomic.AddInt32(&s.curCount,int32(count))
        return
    }
    atomic.StoreInt64(&s.unixSecond,sec)
    atomic.StoreInt32(&s.curCount, int32(count))
}

func (s *SecondLimit) Wait()bool {
    for {
        sec := time.Now().Unix()
        if (sec == atomic.LoadInt64(&s.unixSecond)) && s.curCount == s.limit {
            time.Sleep(time.Microsecond)
            logs.Debug("limit is running,limit:%d s.curCount:%d",s.limit,s.curCount)
            continue
        }

        if sec != atomic.LoadInt64(&s.unixSecond) {
            atomic.StoreInt64(&s.unixSecond,sec)
            atomic.StoreInt32(&s.curCount,0)
        }
        logs.Debug("limit is exited")
        return false
    }
}

Summary

This time, the first half of the log collection is basically processed, and the log will be thrown into es later and rendered on the page

Keywords: Go github Linux Unix Zookeeper

Added by laflair13 on Tue, 14 May 2019 23:30:40 +0300

Programming VIP