duanhui7329 2017-01-09 07:47
浏览 89
已采纳

conn.flush()不会将所有记录刷新到redis

this is code

 func main() {
   ...
   pool := createPool(*redis_server, *redis_pass)
   defer pool.Close()
   c := pool.Get() 
   var i int64
   st := tickSec()
   for i = 0; i < *total; i++ {
      r := time.Now().Unix() - rand.Int63n(60*60*24*31*12)
      score, _ := strconv.Atoi(time.Unix(r, 0).Format("2006010215"))
      id := utee.PlainMd5(uuid.NewUUID().String())
      c.Send("ZADD", "app_a_5512", score, id)
      if i%10000 == 0 {
          c.Flush()
          log.Println("current sync to redis", i)
      }
  }
  //c.Flush()
  c.Close()
  ...
}

if i use c.Close(),the total set 100000,the real sortedset count 100000. but if i use c.Flush(),the total also set 100000, the real sortedset count less than 100000(96932);if i use time.Sleep() in the end of the main func,the total is 100000 too.

when main func exit,the flush func is not complete?and why? thank you!

  • 写回答

1条回答 默认 最新

  • duangewu5234 2017-01-09 14:31
    关注

    The reason that the program works when Close() is called after the loop is that the pooled connection's Close() method reads and discards all pending responses.

    The application should Receive the responses for all commands instead of letting the respones backup and consume memory on the server. There's no need to flush in the loop.

    go func() {
      for i = 0; i < *total; i++ {
         r := time.Now().Unix() - rand.Int63n(60*60*24*31*12)
         score, _ := strconv.Atoi(time.Unix(r, 0).Format("2006010215"))
         id := utee.PlainMd5(uuid.NewUUID().String())
         c.Send("ZADD", "app_a_5512", score, id)
      }
      c.Flush()
    }
    
    for i = 0; i < *total; i++ {
       c.Receive()
    }
    c.Close()
    

    Also, the application should check and handle the errors returns from Send, Flush and Receive.

    本回答被题主选为最佳回答 , 对您是否有帮助呢?
    评论

报告相同问题?

悬赏问题

  • ¥15 Vue3 大型图片数据拖动排序
  • ¥15 划分vlan后不通了
  • ¥15 GDI处理通道视频时总是带有白色锯齿
  • ¥20 用雷电模拟器安装百达屋apk一直闪退
  • ¥15 算能科技20240506咨询(拒绝大模型回答)
  • ¥15 自适应 AR 模型 参数估计Matlab程序
  • ¥100 角动量包络面如何用MATLAB绘制
  • ¥15 merge函数占用内存过大
  • ¥15 使用EMD去噪处理RML2016数据集时候的原理
  • ¥15 神经网络预测均方误差很小 但是图像上看着差别太大