Fork me on GitHub

TDengine时间序列数据库压力测试

发表于 | 更新于 | 分类于 | | 阅读次数：
本文字数： 45k | 阅读时长 ≈ 41 分钟

目录

  • 背景

  • 第一部分 部署压测工具vegeta

  • 第二部分 部署TDengine数据库

  • 第三部分 TDengine数据库RESTful接口介绍

  • 第四部分 压力测试

  • 第五部分 压力测试结果

  • 参考文献及资料

背景

最近涛思数据团队开源了自己的时序数据库，关注度较高。按照官方介绍，性能较为强悍，所以使用压测工具对该数据库进行了性能压测。压测使用的工具Vegeta 是一个用 Go 语言编写的多功能的 HTTP 负载测试工具。Vegeta 提供了命令行工具和一个开发库。

第一部分 部署压测工具vegeta

1.1 部署

从github上下载vegeta安装介质：

1
$ wget https://github.com/tsenart/vegeta/releases/download/cli%2Fv12.5.1/vegeta-12.5.1-linux-amd64.tar.gz

该工具开箱即用，解压tar包：

1
$ tar -zxvf vegeta-12.5.1-linux-amd64.tar.gz

1.2 简单测试使用

我们使用vegeta测试一下下面的压力场景（对百度主页发起每秒100次（-rate=100）的请求，持续10秒（-duration=10s）），测试结果重定向到文件（result/results.bin）：

1
$ echo "GET https://www.baidu.com" |./vegeta attack -duration=10s -rate=100 >result/results.bin

查看一下测试结果：

1
2
3
4
5
6
7
8
9
root@deeplearning:/data/vegeta# ./vegeta report result/results.bin
Requests      [total, rate]            1000, 100.10
Duration      [total, attack, wait]    10.002241136s, 9.990128173s, 12.112963ms
Latencies     [mean, 50, 95, 99, max]  2.315958984s, 1.889487082s, 6.064678156s, 6.583899297s, 6.961230585s
Bytes In      [total, mean]            227000, 227.00
Bytes Out     [total, mean]            0, 0.00
Success       [ratio]                  100.00%
Status Codes  [code:count]             200:1000  
Error Set:

另外可以生成html文件报告（可视化）：

1
root@deeplearning:/data/vegeta# ./vegeta plot result/results.bin > result/plot.html

第二部分 部署TDengine数据库

2.1 制作docker镜像

由于墙的原因我们在VPS上打包镜像，然后本机拉取部署。

为了保证测试环境的隔离性，我们制作docker镜像，使用docker环境进行测试。首先拉取ubuntu的基础镜像：

1
$ docker run -t -i ubuntu:16.04 /bin/bash

通过TDengine源码安装。在这过程有大量操作系统工具未安装，需要使用atp-get安装部署。

2.1.1 第一步 clone项目

1
$ git clone https://github.com/taosdata/TDengine.git

2.1.2 第二步 编译

1
2
$ mkdir build && cd build
$ cmake .. && cmake --build .

2.1.3 第三步 安装

1
$ make install

2.2 生成镜像

打包成镜像，推送到Docker Hub：

1
2
3
root@vultr:~# docker commit a35c43242a8e rongxiang1986/tdengine
root@vultr:~# docker tag rongxiang1986/tdengine rongxiang1986/tdengine:1.0
root@vultr:~# docker push rongxiang1986/tdengine:1.0

2.3 拉取镜像部署

拉取镜像：

1
root@deeplearning:/data/TDengine/TDengine# docker pull rongxiang1986/tdengine:1.0

启动一个docker容器：

1
root@deeplearning:/data/TDengine/TDengine# docker run -t -i --name tdengine -d -p 6020:6020 rongxiang1986/tdengine:1.0 /bin/bash

查看正在运行的容器：

1
2
root@deeplearning:/data/TDengine/TDengine# docker ps
# bec2e166c29f

进入容器：

1
root@deeplearning:/data/TDengine/TDengine# docker attach bec2e166c29f

最后启动数据库服务，下面是启动回显信息：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
root@bec2e166c29f:~/TDengine/TDengine/build# ./build/bin/taosd -c test/cfg &
[1] 23
root@bec2e166c29f:~/TDengine/TDengine/build# TDengine:[23]: Starting TDengine service...
07/17 14:48:21.615248 7ff90254e700 UTL timezone not configured, set to system default:Etc/UTC (Etc, +0000)
07/17 14:48:21.615261 7ff90254e700 UTL locale not configured, set to system default:C
07/17 14:48:21.616266 7ff90254e700 UTL    taos config & system info:
07/17 14:48:21.616274 7ff90254e700 UTL ==================================
07/17 14:48:21.616278 7ff90254e700 UTL  internalIp:             172.17.0.3 
07/17 14:48:21.616283 7ff90254e700 UTL  localIp:                172.17.0.3 
07/17 14:48:21.616288 7ff90254e700 UTL  httpIp:                 0.0.0.0 
07/17 14:48:21.616294 7ff90254e700 UTL  httpPort:               6020 
07/17 14:48:21.616299 7ff90254e700 UTL  mgmtShellPort:          6030 
07/17 14:48:21.616304 7ff90254e700 UTL  vnodeShellPort:         6035 
07/17 14:48:21.616308 7ff90254e700 UTL  configDir:              test/cfg 
07/17 14:48:21.616312 7ff90254e700 UTL  dataDir:                /root/TDengine/TDengine/build/test/data 
07/17 14:48:21.616318 7ff90254e700 UTL  logDir:                 /root/TDengine/TDengine/build/test/log 
07/17 14:48:21.616325 7ff90254e700 UTL  scriptDir:              /etc/taos 
07/17 14:48:21.616330 7ff90254e700 UTL  numOfThreadsPerCore:    1.000000 
07/17 14:48:21.616338 7ff90254e700 UTL  ratioOfQueryThreads:    0.500000 
07/17 14:48:21.616344 7ff90254e700 UTL  numOfVnodesPerCore:     8 
07/17 14:48:21.616349 7ff90254e700 UTL  numOfTotalVnodes:       0 
07/17 14:48:21.616355 7ff90254e700 UTL  tables:                 1000 
07/17 14:48:21.616360 7ff90254e700 UTL  cache:                  16384(byte)
07/17 14:48:21.616366 7ff90254e700 UTL  rows:                   4096 
07/17 14:48:21.616371 7ff90254e700 UTL  fileBlockMinPercent:    0.250000 
07/17 14:48:21.616376 7ff90254e700 UTL  ablocks:                4 
07/17 14:48:21.616382 7ff90254e700 UTL  tblocks:                100 
07/17 14:48:21.616386 7ff90254e700 UTL  monitorInterval:        30(s)
07/17 14:48:21.616391 7ff90254e700 UTL  rpcTimer:               300(ms)
07/17 14:48:21.616396 7ff90254e700 UTL  rpcMaxTime:             600(s)
07/17 14:48:21.616402 7ff90254e700 UTL  ctime:                  3600(s)
07/17 14:48:21.616407 7ff90254e700 UTL  statusInterval:         1(s)
07/17 14:48:21.616414 7ff90254e700 UTL  shellActivityTimer:     3(s)
07/17 14:48:21.616420 7ff90254e700 UTL  meterMetaKeepTimer:     7200(s)
07/17 14:48:21.616423 7ff90254e700 UTL  metricMetaKeepTimer:    600(s)
07/17 14:48:21.616428 7ff90254e700 UTL  maxUsers:               1000 
07/17 14:48:21.616432 7ff90254e700 UTL  maxDbs:                 1000 
07/17 14:48:21.616439 7ff90254e700 UTL  maxTables:              650000 
07/17 14:48:21.616442 7ff90254e700 UTL  maxVGroups:             1000 
07/17 14:48:21.616445 7ff90254e700 UTL  minSlidingTime:         10(ms)
07/17 14:48:21.616454 7ff90254e700 UTL  minIntervalTime:        10(ms)
07/17 14:48:21.616460 7ff90254e700 UTL  maxStreamCompDelay:     20000(ms)
07/17 14:48:21.616463 7ff90254e700 UTL  maxFirstStreamCompDelay:10000(ms)
07/17 14:48:21.616467 7ff90254e700 UTL  retryStreamCompDelay:   10(ms)
07/17 14:48:21.616470 7ff90254e700 UTL  clog:                   1 
07/17 14:48:21.616474 7ff90254e700 UTL  comp:                   2 
07/17 14:48:21.616480 7ff90254e700 UTL  days:                   10 
07/17 14:48:21.616483 7ff90254e700 UTL  keep:                   3650 
07/17 14:48:21.616488 7ff90254e700 UTL  defaultDB:               
07/17 14:48:21.616494 7ff90254e700 UTL  defaultUser:            root 
07/17 14:48:21.616503 7ff90254e700 UTL  defaultPass:            taosdata 
07/17 14:48:21.616508 7ff90254e700 UTL  timezone:               Etc/UTC (Etc, +0000) 
07/17 14:48:21.616515 7ff90254e700 UTL  locale:                 C 
07/17 14:48:21.616520 7ff90254e700 UTL  charset:                UTF-8 
07/17 14:48:21.616525 7ff90254e700 UTL  maxShellConns:          2000 
07/17 14:48:21.616531 7ff90254e700 UTL  maxMeterConnections:    10000 
07/17 14:48:21.616535 7ff90254e700 UTL  maxMgmtConnections:     2000 
07/17 14:48:21.616540 7ff90254e700 UTL  maxVnodeConnections:    10000 
07/17 14:48:21.616543 7ff90254e700 UTL  enableHttp:             1 
07/17 14:48:21.616549 7ff90254e700 UTL  enableMonitor:          1 
07/17 14:48:21.616553 7ff90254e700 UTL  httpCacheSessions:      2000 
07/17 14:48:21.616556 7ff90254e700 UTL  httpMaxThreads:         2 
07/17 14:48:21.616561 7ff90254e700 UTL  numOfLogLines:          10000000 
07/17 14:48:21.616565 7ff90254e700 UTL  asyncLog:               1 
07/17 14:48:21.616570 7ff90254e700 UTL  debugFlag:              131 
07/17 14:48:21.616574 7ff90254e700 UTL  mDebugFlag:             135 
07/17 14:48:21.616579 7ff90254e700 UTL  dDebugFlag:             131 
07/17 14:48:21.616584 7ff90254e700 UTL  sdbDebugFlag:           135 
07/17 14:48:21.616590 7ff90254e700 UTL  taosDebugFlag:          131 
07/17 14:48:21.616595 7ff90254e700 UTL  tmrDebugFlag:           131 
07/17 14:48:21.616601 7ff90254e700 UTL  cDebugFlag:             131 
07/17 14:48:21.616605 7ff90254e700 UTL  jniDebugFlag:           131 
07/17 14:48:21.616612 7ff90254e700 UTL  odbcDebugFlag:          131 
07/17 14:48:21.616617 7ff90254e700 UTL  uDebugFlag:             131 
07/17 14:48:21.616621 7ff90254e700 UTL  httpDebugFlag:          131 
07/17 14:48:21.616629 7ff90254e700 UTL  monitorDebugFlag:       131 
07/17 14:48:21.616634 7ff90254e700 UTL  qDebugFlag:             131 
07/17 14:48:21.616637 7ff90254e700 UTL  gitinfo:                82cbce3261d06ab37c3bd4786c7b2e3d2316c42a 
07/17 14:48:21.616643 7ff90254e700 UTL  buildinfo:              Built by ubuntu at 2019-07-05 18:42 
07/17 14:48:21.616648 7ff90254e700 UTL  version:                1.6.0.0 
07/17 14:48:21.616653 7ff90254e700 UTL  os pageSize:            4096(KB)
07/17 14:48:21.616658 7ff90254e700 UTL  os openMax:             1048576
07/17 14:48:21.616662 7ff90254e700 UTL  os streamMax:           16
07/17 14:48:21.616666 7ff90254e700 UTL  os numOfCores:          8
07/17 14:48:21.616670 7ff90254e700 UTL  os totalDisk:           426(GB)
07/17 14:48:21.616676 7ff90254e700 UTL  os totalMemory:         32028(MB)
07/17 14:48:21.616682 7ff90254e700 UTL  os sysname:             Linux
07/17 14:48:21.616686 7ff90254e700 UTL  os nodename:            bec2e166c29f
07/17 14:48:21.616690 7ff90254e700 UTL  os release:             4.15.0-51-generic
07/17 14:48:21.616694 7ff90254e700 UTL  os version:             #55~16.04.1-Ubuntu SMP Thu May 16 09:24:37 UTC 2019
07/17 14:48:21.616700 7ff90254e700 UTL  os machine:             x86_64
07/17 14:48:21.616707 7ff90254e700 UTL ==================================
07/17 14:48:21.616712 7ff90254e700 DND Server IP address is:172.17.0.3
07/17 14:48:21.616717 7ff90254e700 DND starting to initialize TDengine engine ...
07/17 14:48:21.619440 7ff90254e700 HTP failed to open telegraf schema config file:test/cfg/taos.telegraf.cfg, use default schema
07/17 14:48:21.869736 7ff90254e700 DND vnode is initialized successfully
07/17 14:48:21.869780 7ff90254e700 MND starting to initialize TDengine mgmt ...
07/17 14:48:21.872494 7ff90254e700 MND first access, set total vnodes:64
07/17 14:48:21.917758 7ff90254e700 MND TDengine mgmt is initialized successfully
07/17 14:48:21.917781 7ff90254e700 HTP starting to initialize http service ...
07/17 14:48:21.918417 7ff90254e700 DND TDengine is initialized successfully
07/17 14:48:21.918533 7ff8e4f41700 HTP http service init success at ip:0.0.0.0:6020
TDengine:[23]: Started TDengine service successfully.
07/17 14:48:22.022278 7ff8ff835700 MON starting to initialize monitor service ..
07/17 14:48:22.022747 7ff8eb109700 MND user:monitor login from 172.17.0.3, code:0
07/17 14:48:22.024412 7ff901038700 MON dnode:172.17.0.3 is started
07/17 14:48:22.026780 7ff901038700 MON monitor service init success

上面回显service init success说明服务service启动成功。

2.4 镜像使用说明

TDengine项目地址：https://github.com/taosdata/TDengine。使用ubuntu16.04作为基础镜像，部署安装TDengine。拉取镜像后，启动容器后：

  • 启动服务：To start the TDengine server, run the command below in terminal:

    1
    $ /root/TDengine/TDengine/build# ./build/bin/taosd -c test/cfg

  • 启动客户端：In another terminal, use the TDengine shell to connect the server:

    1
    $ /root/TDengine/TDengine/build#./build/bin/taos -c test/cfg

第三部分 TDengine数据库RESTful接口介绍

按照官方给的例子我们新建案例数据库和表，并且新增数据记录。首先进入shell交互：

1
2
3
4
5
6
7
root@a35c43242a8e:~/TDengine/TDengine/build# ./build/bin/taos -c test/cfg
07/17 04:30:19.818000 7fa154b0e700 MND user:root login from 172.17.0.2, code:0

Welcome to the TDengine shell, server version:1.6.0.0  client version:1.6.0.0
Copyright (c) 2017 by TAOS Data, Inc. All rights reserved.

taos>

创建案例数据：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
taos> create database db;
07/17 04:31:46.970580 7fa14ffff700 MND DB:0.db is created by root
Query OK, 1 row(s) affected (0.001848s)

taos> use db;
Database changed.

taos> create table t (ts timestamp, cdata int);
Query OK, 1 row(s) affected (0.334998s)

taos> insert into t values ('2019-07-15 10:00:00', 10);
Query OK, 1 row(s) affected (0.001639s)

taos> insert into t values ('2019-07-15 10:01:05', 20);
Query OK, 1 row(s) affected (0.000245s)

taos> select * from t;
          ts          |   cdata   |
===================================
 19-07-15 10:00:00.000|         10|
 19-07-15 10:01:05.000|         20|
Query OK, 2 row(s) in set (0.001408s)

退出shell交互后，我们使用Restfull接口与数据库交互。

注意：目前RESTfull接口认证方式使用Http Basic Authorization请求格式，token使用base64(username:password)，即base64(root:taosdata)=cm9vdDp0YW9zZGF0YQ==

可以在在线网站上：https://www.base64encode.org/ encode一下。

1
2
3
4
root@a35c43242a8e:~/TDengine/TDengine/build# curl -H 'Authorization: Basic cm9vdDp0YW9zZGF0YQ==' -d 'select * from db.t' localhost:6020/rest/sql
#下面是回显：
07/17 04:33:34.834283 7fa154b0e700 MND user:root login from 172.17.0.2, code:0
{"status":"succ","head":["ts","cdata"],"data":[["2019-07-15 10:00:00.000",10],["2019-07-15 10:01:05.000",20]],"rows":2}

返回结果是一个JSON格式串，规范化一下：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
{  
   "status":"succ",
   "head":[  
      "ts",
      "cdata"
   ],
   "data":[  
      [  
         "2019-07-15 10:00:00.000",
         10
      ],
      [  
         "2019-07-15 10:01:05.000",
         20
      ]
   ],
   "rows":2
}

第四部分 对RESTful接口压力测试

最后我们使用vegeta对restful接口进行压力测试：

4.1 查询压测

使用目标文件的内容进行压力测试。

首先创建target.txt文件以及数据文件data.json，内容如下：

1
2
3
4
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# cat target.txt
POST http://localhost:6020/rest/sql
Authorization: Basic cm9vdDp0YW9zZGF0YQ==
@data.json
1
2
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# cat data.json
select * from db.t

最后使用下面的命令开始压测（每秒15000次请求，持续5分钟）：

1
2
3
4
5
6
7
8
9
10
11
12
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest#../vegeta attack -rate 15000 -targets target.txt -duration 5m > out.dat
# 生成报告
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# ../vegeta report out.dat
# 回显：
Requests      [total, rate]            4500044, 15000.15
Duration      [total, attack, wait]    5m0.000202809s, 4m59.999911126s, 291.683µs
Latencies     [mean, 50, 95, 99, max]  314.712µs, 230.113µs, 869.502µs, 1.501018ms, 205.870168ms
Bytes In      [total, mean]            954009328, 212.00
Bytes Out     [total, mean]            90000880, 20.00
Success       [ratio]                  100.00%
Status Codes  [code:count]             200:4500044  
Error Set:

4.2 写入压测

首先创建writetarget.txt和数据文件：writedata.json，内容如下：

1
2
3
4
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# cat writetarget.txt
POST http://localhost:6020/rest/sql
Authorization: Basic cm9vdDp0YW9zZGF0YQ==
@writedata.json
1
2
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# cat writedata.json
insert into db.cpu values (NOW,20,12);

这里写入语句使用时间函数NOW，保证写入时间无重复。

最后使用命令开始压测（每秒30000次请求，持续1分钟）：

1
2
3
4
5
6
7
8
9
10
11
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# ../vegeta attack -rate 30000 -targets writetarget.txt -duration 1m > writeout.dat
# 生成报告
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# ../vegeta report writeout30000.dat
Requests      [total, rate]            1800018, 30000.31
Duration      [total, attack, wait]    1m0.000154842s, 59.99998565s, 169.192µs
Latencies     [mean, 50, 95, 99, max]  254.824µs, 137.651µs, 904.696µs, 1.687601ms, 12.528831ms
Bytes In      [total, mean]            115201152, 64.00
Bytes Out     [total, mean]            70200702, 39.00
Success       [ratio]                  100.00%
Status Codes  [code:count]             200:1800018  
Error Set:

第五部分 压力测试结果

由于机器环境的差异，只是做了尝试性测试，不代表产品的实际性能。

官方测试报告参考：https://www.taosdata.com/downloads/TDengine_Testing_Report_cn.pdf

5.1 查询

对于查询性能我们每秒15000的请求结果如下：

1
2
3
4
5
6
7
8
9
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# ../vegeta report out15000.dat
Requests      [total, rate]            4500044, 15000.15
Duration      [total, attack, wait]    5m0.000202809s, 4m59.999911126s, 291.683µs
Latencies     [mean, 50, 95, 99, max]  314.712µs, 230.113µs, 869.502µs, 1.501018ms, 205.870168ms
Bytes In      [total, mean]            954009328, 212.00
Bytes Out     [total, mean]            90000880, 20.00
Success       [ratio]                  100.00%
Status Codes  [code:count]             200:4500044  
Error Set:

当提高到每秒20000次时，数据库出现响应失败，成功率只有38.78%：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# ../vegeta report out20000.dat     
Requests      [total, rate]            904794, 2824.86
Duration      [total, attack, wait]    6m19.120695351s, 5m20.296894938s, 58.823800413s
Latencies     [mean, 50, 95, 99, max]  38.726984614s, 43.82040722s, 1m13.377381823s, 1m25.843611324s, 1m52.393219406s
Bytes In      [total, mean]            58181450, 64.30
Bytes Out     [total, mean]            8016960, 8.86
Success       [ratio]                  38.78%
Status Codes  [code:count]             0:503946  200:350896  400:49952  
Error Set:
400 Bad Request
Post http://localhost:6020/rest/sql: dial tcp 0.0.0.0:0->127.0.0.1:6020: bind: address already in use
Post http://localhost:6020/rest/sql: net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Post http://localhost:6020/rest/sql: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
Post http://localhost:6020/rest/sql: dial tcp 0.0.0.0:0->127.0.0.1:6020: bind: address already in use (Client.Timeout exceeded while awaiting headers)
Post http://localhost:6020/rest/sql: read tcp 127.0.0.1:34037->127.0.0.1:6020: read: connection reset by peer
Post http://localhost:6020/rest/sql: read tcp 127.0.0.1:53020->127.0.0.1:6020: read: connection reset by peer
Post http://localhost:6020/rest/sql: read tcp 127.0.0.1:47081->127.0.0.1:6020: read: connection reset by peer
Post http://localhost:6020/rest/sql: read tcp 127.0.0.1:35442->127.0.0.1:6020: read: connection reset by peer
Post http://localhost:6020/rest/sql: read tcp 127.0.0.1:58175->127.0.0.1:6020: read: connection reset by peer
Post http://localhost:6020/rest/sql: read tcp 127.0.0.1:41857->127.0.0.1:6020: read: connection reset by peer
Post http://localhost:6020/rest/sql: read tcp 127.0.0.1:58943->127.0.0.1:6020: read: connection reset by peer

5.2 写入

对于写入测试。对于查询性能我们每秒30000的请求结果如下：

1
2
3
4
5
6
7
8
9
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# ../vegeta report writeout30000.dat
Requests      [total, rate]            1800018, 30000.31
Duration      [total, attack, wait]    1m0.000154842s, 59.99998565s, 169.192µs
Latencies     [mean, 50, 95, 99, max]  254.824µs, 137.651µs, 904.696µs, 1.687601ms, 12.528831ms
Bytes In      [total, mean]            115201152, 64.00
Bytes Out     [total, mean]            70200702, 39.00
Success       [ratio]                  100.00%
Status Codes  [code:count]             200:1800018  
Error Set:

当提高到每秒50000次时，数据库性能开始恶化，成功率只有55.73%：

1
2
3
4
5
6
7
8
9
10
11
12
13
root@bec2e166c29f:~/TDengine/vegeta/TDengineTest# ../vegeta report writeout50000.dat
Requests      [total, rate]            160712, 2676.70
Duration      [total, attack, wait]    1m47.789030336s, 1m0.041100866s, 47.74792947s
Latencies     [mean, 50, 95, 99, max]  36.064568863s, 39.103737526s, 50.372069878s, 54.866214748s, 1m12.470342472s
Bytes In      [total, mean]            5731840, 35.67
Bytes Out     [total, mean]            3492840, 21.73
Success       [ratio]                  55.73%
Status Codes  [code:count]             0:71152  200:89560  
Error Set:
Post http://localhost:6020/rest/sql: dial tcp 0.0.0.0:0->127.0.0.1:6020: bind: address already in use
Post http://localhost:6020/rest/sql: net/http: request canceled (Client.Timeout exceeded while awaiting headers)
Post http://localhost:6020/rest/sql: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
Post http://localhost:6020/rest/sql: dial tcp 0.0.0.0:0->127.0.0.1:6020: bind: address already in use (Client.Timeout exceeded while awaiting headers)

结论：单机环境下同等场景（都是压测RESTfull接口）下，TDengine可以抗住每秒15000次的读请求和每秒30000次写请求。influxdb只能抗住每秒6000次持续读、写。按照官网介绍如果写入是批量形式会更快。

参考文献和材料

1、推荐一款高性能 HTTP 负载测试工具 Vegeta 链接：https://www.hi-linux.com/posts/4650.html

2、TDengine官网 链接：https://www.taosdata.com/cn/

3、比Hadoop快至少10倍的物联网大数据平台，我把它开源了 链接：https://weibo.com/ttarticle/p/show?id=2309404394278649462890

本文标题:TDengine时间序列数据库压力测试

文章作者:rong xiang

发布时间:2019年07月01日 - 23:07

最后更新:2022年10月25日 - 23:10

原始链接:https://zjrongxiang.github.io/posts/eb6b6c24/

许可协议: 署名-非商业性使用-禁止演绎 4.0 国际 转载请保留原文链接及作者。

0%