（六）使用ab（Apache-BenchMark）壓測 Dubbo

阿新 • • 發佈：2022-03-11

1、準備

使用ab（Apache-BenchMark）壓測consumer的Controller，觸發呼叫。

dubbo的consumer發起呼叫，呼叫provider暴露的介面。

provider做1w次迴圈，生成隨機數做累加。

2、ab壓測情況

2.1、1k 資料

情況一：

三臺同機房的機器，利用兩臺consumer請求同一臺provider

機器：

consumer1  2h4g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer2  4h8g   -server -Xmx4g -Xms4g -XX:+UseG1GC
provider   2h4g	  -server -Xmx2g -Xms2g -XX:+UseG1GC

因為單機無法讓provider機器cpu壓滿

取10併發，consumer2 進行100w次請求，consumer1 進行 80w次（為了和consumer2 在同一時間完成）請求。

 ab -n 1000000 -c  10 http://127.0.0.1:8091/consumer/stressTest/string1k

provider情況：

top - 09:51:00 up 1137 days, 18:04,  4 users,  load average: 7.07, 3.55, 2.35
Tasks:   1 total,   0 running,   1 sleeping,   0 stopped,   0 zombie
Cpu0  : 80.4%us, 11.0%sy,  0.0%ni,  1.0%id,  0.0%wa,  0.0%hi,  7.6%si,  0.0%st
Cpu1  : 88.6%us,  9.4%sy,  0.0%ni,  2.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   3924680k total,  3124704k used,   799976k free,   102508k buffers
Swap:  2097144k total,     6156k used,  2090988k free,  1074068k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                                                                                       
 1089 root      20   0 4905m 1.5g  13m S 196.8 39.3  49:05.17 java -Dfile.encoding=utf-8 -server -Xmx2g -Xms2g -XX:+UseG1GC -Xloggc:/data/dubboStress/logs/dubbo_gc.log -XX

可以看到CPU負載過高。

consumer1、consumer2 幾乎同時執行結束。

consumer1 2h4g 結果：

Concurrency Level:      10
Time taken for tests:   197.835 seconds
Complete requests:      800000
Failed requests:        0
Write errors:           0
Total transferred:      87200000 bytes
HTML transferred:       3200000 bytes
Requests per second:    4043.77 [#/sec] (mean)
Time per request:       2.473 [ms] (mean)
Time per request:       0.247 [ms] (mean, across all concurrent requests)
Transfer rate:          430.44 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%      2
  66%      3
  75%      3
  80%      3
  90%      4
  95%      4
  98%      5
  99%      6
 100%    224 (longest request)

consumer2 4h8g：

Concurrency Level:      10
Time taken for tests:   220.944 seconds
Complete requests:      1000000
Failed requests:        0
Write errors:           0
Total transferred:      109000000 bytes
HTML transferred:       4000000 bytes
Requests per second:    4526.03 [#/sec] (mean)
Time per request:       2.209 [ms] (mean)
Time per request:       0.221 [ms] (mean, across all concurrent requests)
Transfer rate:          481.77 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%      2
  66%      2
  75%      3
  80%      3
  90%      3
  95%      4
  98%      4
  99%      5
 100%    261 (longest request)

可以看到兩臺機tps加起來有 8500多，95%響應時間都在 5ms以內。

情況二：

我再加一臺跨機房的機器測試一下。

consumer1  2h4g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer2  4h8g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer3  2h4g   -server -Xmx2g -Xms2g -XX:+UseG1GC （跨網段）
provider   2h4g	  -server -Xmx2g -Xms2g -XX:+UseG1GC

consumer1：

Concurrency Level:      10
Time taken for tests:   269.711 seconds
Complete requests:      1000000
Failed requests:        0
Write errors:           0
Total transferred:      109000000 bytes
HTML transferred:       4000000 bytes
Requests per second:    3707.68 [#/sec] (mean)
Time per request:       2.697 [ms] (mean)
Time per request:       0.270 [ms] (mean, across all concurrent requests)
Transfer rate:          394.66 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%      2
  66%      3
  75%      3
  80%      3
  90%      4
  95%      5
  98%      5
  99%      6
 100%    233 (longest request)

consumer2：

Concurrency Level:      10
Time taken for tests:   251.411 seconds
Complete requests:      1000000
Failed requests:        0
Write errors:           0
Total transferred:      109000000 bytes
HTML transferred:       4000000 bytes
Requests per second:    3977.55 [#/sec] (mean)
Time per request:       2.514 [ms] (mean)
Time per request:       0.251 [ms] (mean, across all concurrent requests)
Transfer rate:          423.39 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.1      0       2
Processing:     1    2   2.8      2     201
Waiting:        0    2   2.8      2     201
Total:          1    2   2.8      2     201

Percentage of the requests served within a certain time (ms)
  50%      2
  66%      3
  75%      3
  80%      3
  90%      4
  95%      4
  98%      5
  99%      6
 100%    201 (longest request)

consumer3：

（跨機房，RTT較慢）

Concurrency Level:      10
Time taken for tests:   253.211 seconds
Complete requests:      220941
Failed requests:        0
Write errors:           0
Total transferred:      24082569 bytes
HTML transferred:       883764 bytes
Requests per second:    872.56 [#/sec] (mean)
Time per request:       11.461 [ms] (mean)
Time per request:       1.146 [ms] (mean, across all concurrent requests)
Transfer rate:          92.88 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%     11
  66%     11
  75%     12
  80%     12
  90%     13
  95%     14
  98%     15
  99%     17
 100%    281 (longest request)

可以看到tps累加也差不多是 8500多。

情況三：

換成都是同機房的機器。

consumer1  2h4g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer2  4h8g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer3  8h8g   -server -Xmx4g -Xms4g -XX:+UseG1GC
provider   2h4g	  -server -Xmx2g -Xms2g -XX:+UseG1GC

結果差不多，TPS有增加，誤差來自於 consumer1執行完成，consumer2、consumer3還沒執行完成，但是時間差在5秒以內。

consumer1：

Concurrency Level:      10
Time taken for tests:   299.224 seconds
Complete requests:      1000000
Requests per second:    3341.98 [#/sec] (mean)
Transfer rate:          355.74 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%      3
  66%      3
  75%      3
  80%      4
  90%      4
  95%      5
  98%      6
  99%      7

consmer2：

Concurrency Level:      10
Time taken for tests:   339.512 seconds
Complete requests:      1000000
Requests per second:    2945.41 [#/sec] (mean)
Transfer rate:          313.52 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%      3
  66%      4
  75%      4
  80%      4
  90%      5
  95%      6
  98%      7
  99%      8

consmer3：

Concurrency Level:      10
Time taken for tests:   368.969 seconds
Complete requests:      1000000
Requests per second:    2710.26 [#/sec] (mean)
Transfer rate:          288.49 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%      3
  66%      4
  75%      4
  80%      5
  90%      5
  95%      6
  98%      7
  99%      8

TPS差不多近9000，RTT響應時間差不多。

provider 伺服器情況（CPU爆高）：

top - 17:53:54 up 1141 days,  2:07,  2 users,  load average: 7.14, 5.33, 4.19
Tasks:   1 total,   0 running,   1 sleeping,   0 stopped,   0 zombie
Cpu0  : 84.3%us,  8.7%sy,  0.0%ni,  0.7%id,  0.0%wa,  0.0%hi,  6.4%si,  0.0%st
Cpu1  : 89.4%us,  9.6%sy,  0.0%ni,  1.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   3924680k total,  3889636k used,    35044k free,   135212k buffers
Swap:  2097144k total,     5988k used,  2091156k free,  1784112k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                                                                                       
 1089 root      20   0 4908m 1.5g  13m S 198.1 39.4 114:41.32 java -Dfile.encoding=utf-8 -server -Xmx2g -Xms2g -XX:+UseG1GC -Xloggc:/data/dubboStress/logs/dubbo_gc.log

2.2、100k資料

情況一：

consumer1  2h4g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer2  8h8g   -server -Xmx4g -Xms4g -XX:+UseG1GC
provider   2h4g	  -server -Xmx2g -Xms2g -XX:+UseG1GC

10w請求，20併發（100k請求太慢了。。）

ab -n 100000 -c 20 http://127.0.0.1:8091/consumer/stressTest/string100k

consumer1：

Concurrency Level:      20
Time taken for tests:   184.131 seconds
Requests per second:    543.09 [#/sec] (mean)
Transfer rate:          57.81 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%     36
  66%     38
  75%     39
  80%     40
  90%     42
  95%     44
  98%     48
  99%     50

consumer2:

Concurrency Level:      20
Time taken for tests:   185.186 seconds
Complete requests:      100000
Requests per second:    540.00 [#/sec] (mean)
Transfer rate:          57.48 [Kbytes/sec] received

Percentage of the requests served within a certain time (ms)
  50%     36
  66%     38
  75%     39
  80%     40
  90%     42
  95%     45
  98%     49
  99%     52

tps差不多1000，RTT 95% 45ms

情況二：

consumer1  2h4g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer2  8h8g   -server -Xmx4g -Xms4g -XX:+UseG1GC
provider   2h4g	  -server -Xmx2g -Xms2g -XX:+UseG1GC

10w請求 10併發

ab -n 100000 -c 10 http://127.0.0.1:8091/consumer/stressTest/string100k

consumer1：

Concurrency Level:      10
Time taken for tests:   184.276 seconds
Complete requests:      100000
Failed requests:        0
Write errors:           0
Total transferred:      10900000 bytes
HTML transferred:       400000 bytes
Requests per second:    542.66 [#/sec] (mean)
Time per request:       18.428 [ms] (mean)
Time per request:       1.843 [ms] (mean, across all concurrent requests)
Transfer rate:          57.76 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.0      0       1
Processing:     7   18   2.8     18      67
Waiting:        7   18   2.8     18      67
Total:          7   18   2.8     18      67

Percentage of the requests served within a certain time (ms)
  50%     18
  66%     19
  75%     20
  80%     20
  90%     21
  95%     23
  98%     25
  99%     28

consumer2：

Concurrency Level:      10
Time taken for tests:   183.730 seconds
Complete requests:      100000
Failed requests:        0
Write errors:           0
Total transferred:      10900000 bytes
HTML transferred:       400000 bytes
Requests per second:    544.28 [#/sec] (mean)
Time per request:       18.373 [ms] (mean)
Time per request:       1.837 [ms] (mean, across all concurrent requests)
Transfer rate:          57.94 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.0      0       1
Processing:     7   18   2.8     18      56
Waiting:        7   18   2.8     18      55
Total:          7   18   2.8     18      56

Percentage of the requests served within a certain time (ms)
  50%     18
  66%     19
  75%     20
  80%     20
  90%     22
  95%     23
  98%     25
  99%     27

TPS:1100 ，RTT 95% 23ms

說明並不是併發數越高越好。

情況三：

加一臺機器。

consumer1  2h4g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer2  4h8g   -server -Xmx4g -Xms4g -XX:+UseG1GC
consumer3  8h8g   -server -Xmx4g -Xms4g -XX:+UseG1GC
provider   2h4g	  -server -Xmx2g -Xms2g -XX:+UseG1GC

10w請求 10併發

ab -n 100000 -c 10 http://127.0.0.1:8091/consumer/stressTest/string100k

consumer1先執行完成，手動停止 2、3 ，存在2s誤差。

consumer1：

Concurrency Level:      10
Time taken for tests:   179.200 seconds
Complete requests:      100000
Failed requests:        0
Write errors:           0
Total transferred:      10900000 bytes
HTML transferred:       400000 bytes
Requests per second:    558.04 [#/sec] (mean)
Time per request:       17.920 [ms] (mean)
Time per request:       1.792 [ms] (mean, across all concurrent requests)
Transfer rate:          59.40 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.0      0       1
Processing:     7   18   4.3     17      55
Waiting:        7   18   4.3     17      54
Total:          7   18   4.3     17      55

Percentage of the requests served within a certain time (ms)
  50%     17
  66%     19
  75%     21
  80%     22
  90%     23
  95%     25
  98%     27
  99%     29

consumer2：

Concurrency Level:      10
Time taken for tests:   190.898 seconds
Complete requests:      73115
Failed requests:        0
Write errors:           0
Total transferred:      7969535 bytes
HTML transferred:       292460 bytes
Requests per second:    383.01 [#/sec] (mean)
Time per request:       26.109 [ms] (mean)
Time per request:       2.611 [ms] (mean, across all concurrent requests)
Transfer rate:          40.77 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.0      0       1
Processing:     7   26   5.6     26      67
Waiting:        7   26   5.6     26      66
Total:          7   26   5.6     26      67

Percentage of the requests served within a certain time (ms)
  50%     26
  66%     28
  75%     29
  80%     30
  90%     33
  95%     35
  98%     38
  99%     41

consumer3：

Concurrency Level:      10
Time taken for tests:   188.066 seconds
Complete requests:      44403
Failed requests:        0
Write errors:           0
Total transferred:      4839927 bytes
HTML transferred:       177612 bytes
Requests per second:    236.10 [#/sec] (mean)
Time per request:       42.354 [ms] (mean)
Time per request:       4.235 [ms] (mean, across all concurrent requests)
Transfer rate:          25.13 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.0      0       1
Processing:    11   42  13.0     42     182
Waiting:       11   42  13.0     42     181
Total:         11   42  13.0     42     182

Percentage of the requests served within a certain time (ms)
  50%     42
  66%     48
  75%     51
  80%     53
  90%     58
  95%     64
  98%     71
  99%     76

tps 1100左右，RTT 要視機器而定。平均就是95% 41ms

provider 伺服器情況：

top - 10:18:21 up 1141 days, 18:31,  3 users,  load average: 5.02, 4.67, 4.14
Tasks:   1 total,   0 running,   1 sleeping,   0 stopped,   0 zombie
Cpu0  :  0.1%us,  0.0%sy,  0.0%ni, 99.9%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Cpu1  :  0.1%us,  0.0%sy,  0.0%ni, 99.8%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   3924680k total,  3892608k used,    32072k free,   141252k buffers
Swap:  2097144k total,     5988k used,  2091156k free,  1726716k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                                                                                       
 1089 root      20   0 4908m 1.5g  13m S 198.5 40.8 188:01.30 java -Dfile.encoding=utf-8 -server -Xmx2g -Xms2g -XX:+UseG1GC

100w 請求 10 併發這裡不放了，TPS差不多，RTT會稍微高一點

3、分析

2.1、dubbo執行緒 WAITING

試過調大了 dubbo （預設是200）的執行緒數，發現效果並沒有顯著的提升。

"DubboServerHandler-172.16.179.198:20890-thread-15" #54 daemon prio=5 os_prio=0 tid=0x00007f32c0015800 nid=0x4a4 waiting on condition [0x00007f32b5046000]
   java.lang.Thread.State: WAITING (parking)
        at sun.misc.Unsafe.park(Native Method)
        - parking to wait for  <0x0000000080e805a8> (a java.util.concurrent.SynchronousQueue$TransferStack)
        at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
        at java.util.concurrent.SynchronousQueue$TransferStack.awaitFulfill(SynchronousQueue.java:458)
        at java.util.concurrent.SynchronousQueue$TransferStack.transfer(SynchronousQueue.java:362)
        at java.util.concurrent.SynchronousQueue.take(SynchronousQueue.java:924)
        at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1074)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1134)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at org.apache.dubbo.common.threadlocal.InternalRunnable.run(InternalRunnable.java:41)
        at java.lang.Thread.run(Thread.java:748)

暫時無解，網上說法是執行緒都在同步佇列，太多反而出現資源等待的情況。

2.2、調大ab的併發數

超過20執行緒，RTT會變得很大，原因：

執行緒的上下文切換嚴重
執行緒等待CPU排程（ab發出請求）

隨後TPS會達到瓶頸，並不是十分顯著。

4、結果彙總

provider機器：

2h4g

CentOS release 6.4 (Final)
model name      : QEMU Virtual CPU version 2.5+
stepping        : 3
cpu MHz         : 2099.998
cache size      : 4096 KB

JVM：

jdk1.8
-server -Xmx2g -Xms2g -XX:+UseG1GC

1、多機，以壓滿CPU為目的

ab -n 1000000 -c  10

詳情見上。

	1k dubbo+hessian2	100k dubbo+hessian2
TPS	8500~9000	1000~1100
RTT	95% 5ms	95% 41ms
OOM	無	無

2、單機，不壓滿CPU

單機：（沒有壓滿CPU，provider機器同上，使用consumer2）

過程省略。

ab -n 1000000 -c  10

	1k dubbo+hessian2	100k dubbo+hessian2
TPS	4000~4500	600~700
RTT	95% 4ms	95% 19ms
OOM	無	無
CPU	125%+	150%

參考：

壓測dubbo：https://blog.csdn.net/u013815546/article/details/101385888
壓測dubbo：（有點不一樣）https://blog.csdn.net/cyjs1988/article/details/84258046
https://blog.51cto.com/ydhome/1861956
官網：https://jmeter-plugins.org/wiki/Start
jmeter使用：https://blog.csdn.net/github_27109687/article/details/71968662
使用 rest 協議：https://dangdangdotcom.github.io/dubbox/rest.html
基於Dubbo的Hessian協議實現遠端呼叫:http://shiyanjun.cn/archives/349.html

Java資源分享：

Java學習路線思維導圖+Java學習視訊+簡歷模板+Java電子書

（六）使用ab（Apache-BenchMark）壓測 Dubbo

1、準備使用ab（Apache-BenchMark）壓測consumer的Controller，觸發呼叫。 dubbo的consumer發起呼叫，呼叫provider暴露的介面。

hive能替代oracle_走近大資料之Hive入門（六、Hive的資料模型）

技術標籤：hive能替代oracle 一、Hive的資料儲存基於HDFS(Hadoop的資料倉庫)沒有專門的資料儲存格式(預設的情況下采用\'\'(製表符)作為列與列之間的分隔符)儲存結構主要包括：資料庫，檔案，表，檢視可以直

（八）JMH的詳細使用，附帶壓測dubbo案例+程式碼

1、JMH簡介 JMH即Java Microbenchmark Harness，是Java用來做基準測試的一個工具，該工具由OpenJDK提供並維護，測試結果可信度高。

（七）使用 Jmeter 壓測Dubbo

壓測思路：壓測consumer的Controller，呼叫provider暴露的介面。 provider做1w次迴圈，生成隨機數做累加。

AB 使用AB對WEB伺服器進行壓測

1、調整系統限制系統預設情況下允許一個程序同時開啟的檔案描述符數量是1024，若我們需要進行超過1024併發測試的話，可能需要調高此值，客戶端和服務端都需要調整。

Apache Flink 進階（六）：Flink 作業執行深度解析

作者：嶽猛整理：毛鶴本文根據 Apache Flink 系列直播課程整理而成，由 Apache Flink Contributor、網易雲音樂實時計算平臺研發工程師嶽猛分享。主要分享內容為 Flink Job 執行作業的流程，文章將從兩個方面進行分

CentOS 7 配置hadoop(六) 配置flume（偽分佈）錯誤: 找不到或無法載入主類 org.apache.flume.tools.GetJavaProperty

一．環境準備：linux + java的Jdk1.6以上版本二．安裝包下載地址：https://flume.apache.org/download.html

【Apache POI】Excel操作（六）：Excel計算公式的讀取和使用

前言在上期：【Apache POI】Excel操作（五）：Excel資料的讀取的部落格中，如果有細心的小夥伴的話，恐怕會發現Excel還有一種資料型別我們並沒有使用到，那就是：HSSFCell.CELL_TYPE_FORMULA 1那這個是啥呢？

OpenGL學習（六）-- 基礎紋理

我的 OpenGL 專題學習目錄，希望和大家一起學習交流進步！ OpenGL學習（一）-- 術語瞭解

Dubbo原始碼解析（六）註冊中心——redis

註冊中心——redis 目標：解釋以為redis實現的註冊中心原理，解讀duubo-registry-redis的原始碼

TiDB Binlog 原始碼閱讀系列文章（六）Pump Storage 介紹（下）

作者：Chunzhu Li 在上篇文章中，我們主要介紹了 Pump Storage 是如何對 binlog 進行持久化儲存、排序、配對的。在文中我們提到 binlog 的持久化鍵值儲存主要是由 valueLog 元件完成的。同時，大家如果在上文點開

認識演演算法之資料結構（六）雜湊表

雜湊表(Hash tables) 什麼是雜湊表在介紹雜湊表的時候，先回顧一下前幾章的內容，再來比較下陣列和連結串列的優缺點。

Flink 系列（六）—— Flink 視窗模型

一、視窗概念在大多數場景下，我們需要統計的資料流都是無界的，因此我們無法等待整個資料流終止後才進行統計。通常情況下，我們只需要對某個時間範圍或者數量範圍內的資料進行統計分析：如每隔五分鐘統計一次過去一

Java SE基礎鞏固（六）：Java IO

到現在為止，Java IO可分為三類：BIO、NIO、AIO。最早出現的是BIO，然後是NIO，最近的是AIO，BIO即Blocking IO，NIO有的文章說是New NIO，也有的文章說是No Blocking IO，我查了一些資料，官網說的應該是No Blocking

go語言20小時從入門到精通（六、工程管理）

在實際的開發工作中，直接呼叫編譯器進行編譯和連結的場景是少而又少，因為在工程中不會簡單到只有一個原始碼檔案，且原始檔之間會有相互的依賴關係。如果這樣一個檔案一個檔案逐步編譯，那不亞於一場災難。 Go語言的

Hive 系列（六）—— Hive 檢視和索引

一、檢視 1.1 簡介 Hive 中的檢視和 RDBMS 中檢視的概念一致，都是一組資料的邏輯表示，本質上就是一條 SELECT 語句的結果集。檢視是純粹的邏輯物件，沒有關聯的儲存 (Hive 3.0.0 引入的物化檢視除外)，當查詢引用檢

HBase 系列（六）——HBase Java API 的基本使用

一、簡述截至到目前 (2019.04)，HBase 有兩個主要的版本，分別是 1.x 和 2.x ，兩個版本的 Java API 有所不同，1.x 中某些方法在 2.x 中被標識為 @deprecated 過時。所以下面關於 API 的樣例，我會分別給出 1.x 和

Hadoop 系列（六）—— HDFS 常用 Shell 命令

1. 顯示當前目錄結構 # 顯示當前目錄結構 hadoop fs -ls<path> # 遞迴顯示當前目錄結構

Scala 系列（六）—— 常用集合型別之 List & Set

一、List字面量 List 是 Scala 中非常重要的一個資料結構，其與 Array(陣列) 非常類似，但是 List 是不可變的，和 Java 中的 List 一樣，其底層實現是連結串列。

Spark 系列（六）—— 累加器與廣播變數

一、簡介在 Spark 中，提供了兩種型別的共享變數：累加器 (accumulator) 與廣播變數 (broadcast variable)：

（六）使用ab（Apache-BenchMark）壓測 Dubbo

1、準備

2、ab壓測情況

2.1、1k 資料

情況一：

情況二：

情況三：

2.2、100k資料

情況一：

情況二：

情況三：

3、分析

2.1、dubbo執行緒 WAITING

2.2、調大ab的併發數

4、結果彙總

1、多機，以壓滿CPU為目的

2、單機，不壓滿CPU

相關推薦