golang value并发安全的另外一种玩法,就是使用atomic.Value,看一段代码。golang
package main import ( "sync" "sync/atomic" "time" ) func main() { var m atomic.Value type Map map[string]string m.Store(make(Map)) var mu sync.Mutex read := func(key string) (val string) { m1 := m.Load().(Map) return m1[key] } insert := func(key, val string) { mu.Lock() defer mu.Unlock() m1 := m.Load().(Map) m2 := make(Map) for k, v := range m1 { m2[k] = v } m2[key] = val m.Store(m2) } go func() { for { insert("k", "v") time.Sleep(100 * time.Millisecond) } }() go func() { for { read("k") } }() time.Sleep(10 * time.Second) }
相对于读写锁,少了一些锁的争抢,不过相对的,带来了一些,内存上的开销,适用于读多写少而且变量占用内存不是特别大的状况,若是用内存存储大量数据,这个并不适合,技术上主要是常见的写时复制(copy-on-write)。安全
另外这个还比较适合程序配置的存储,贴一段官方的栗子架构
var config Value // holds current server configuration // Create initial config value and store into config. config.Store(loadConfig()) go func() { // Reload config every 10 seconds // and update config value with the new version. for { time.Sleep(10 * time.Second) config.Store(loadConfig()) } }() // Create worker goroutines that handle incoming requests // using the latest config value. for i := 0; i < 10; i++ { go func() { for r := range requests() { c := config.Load() // Handle request r using config c. _, _ = r, c } }() }
恩,挺好玩的。并发
更多架构、PHP、GO相关踩坑实践技巧请关注个人公众号:PHP架构师atom