本文最后更新于 1559 天前，其中的信息可能已经有所发展或是发生改变。
[toc]
多年没有折腾过服务器了，直到最近发现曾经在 DigitalOcean 做推广获得的75美元赠金竟然过期了？！再加上自己一直都有重新把博客建起来的想法，便打算入手一台VPS把博客搭起来。因为本人是一位穷逼学生党，自然就只能选择便宜的年付小内存 VPS 。恰逢 pacificrack 这家服务商在搞活动，几美元就可以买一年，赶紧上车。
一、购入
pacificrack 是一家最近才火起来的服务商，据说是国人运营的 quadranet 的子公司(pacificrack 好像就是 quadranet 的曾用名)。机房在 quadranet 自家的机房里，目前走的是 CN2 GT 线路，如果不翻车的话性价比那是非常高了。但是这家的口碑比较两极分化，有的人用起来觉得很不错，有的人却连环翻车，初步估计是分到的母鸡不同的锅，真一切全看脸😂。
购买支持使用支付宝付款，博主购买的是7.25美元一年384MB内存的元旦优惠 VPS ，分分钟付款开机。
二、与 Vultr 的对比测试
由于 Pacificrack 在面板里最高只有 Debian 9 64位版本可选，所有测试都运行在 Debian 9 环境下。同时，为了控制变量，测试用 Vultr VPS 的位置为美西洛杉矶机房，配置为1H1G。
1、服务器性能
IO测试
第一次：
I/O speed(1st run) : 89.4 MB/s
I/O speed(2nd run) : 104 MB/s
I/O speed(3rd run) : 92.0 MB/s
Average I/O speed : 95.1 MB/s
第二次：
I/O speed(1st run) : 90.1 MB/s
I/O speed(2nd run) : 88.0 MB/s
I/O speed(3rd run) : 95.4 MB/s
Average I/O speed : 91.2 MB/s
第三次：
I/O speed(1st run) : 111 MB/s
I/O speed(2nd run) : 112 MB/s
I/O speed(3rd run) : 131 MB/s
Average I/O speed : 118.0 MB/sVultr
I/O Speed( 1.0GB ) : 356 MB/s
I/O Speed( 1.0GB ) : 353 MB/s
I/O Speed( 1.0GB ) : 346 MB/s
Average I/O Speed : 351.7 MB/s
I/O speed(1st run) : 341 MB/s
I/O speed(2nd run) : 351 MB/s
I/O speed(3rd run) : 350 MB/s
Average I/O speed : 347.3 MB/s和 Vultr 相比这样的IO成绩肯定不算快（毕竟人家用的可是 SSD ，但用来建站也差不多够了。
UnixBench跑分测试
BYTE UNIX Benchmarks (Version 5.1.3)
System: Csyvps: GNU/Linux
OS: GNU/Linux -- 4.9.0-4-amd64 -- #1 SMP Debian 4.9.65-3+deb9u1 (2017-12-23)
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: QEMU Virtual CPU version (cpu64-rhel6) (5200.0 bogomips)
x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
01:47:14 up 1:07, 1 user, load average: 0.09, 0.24, 0.30; runlevel 5
------------------------------------------------------------------------
Benchmark Run: Fri Jan 17 2020 01:47:14 - 02:15:24
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 19399515.4 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3052.9 MWIPS (9.6 s, 7 samples)
Execl Throughput 2585.2 lps (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 567586.8 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 138207.8 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1135564.2 KBps (30.0 s, 2 samples)
Pipe Throughput 1376439.2 lps (10.0 s, 7 samples)
Pipe-based Context Switching 134205.9 lps (10.0 s, 7 samples)
Process Creation 6433.5 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 4866.1 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 660.9 lpm (60.1 s, 2 samples)
System Call Overhead 2258575.5 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 19399515.4 1662.3
Double-Precision Whetstone 55.0 3052.9 555.1
Execl Throughput 43.0 2585.2 601.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 567586.8 1433.3
File Copy 256 bufsize 500 maxblocks 1655.0 138207.8 835.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 1135564.2 1957.9
Pipe Throughput 12440.0 1376439.2 1106.5
Pipe-based Context Switching 4000.0 134205.9 335.5
Process Creation 126.0 6433.5 510.6
Shell Scripts (1 concurrent) 42.4 4866.1 1147.7
Shell Scripts (8 concurrent) 6.0 660.9 1101.6
System Call Overhead 15000.0 2258575.5 1505.7
========
System Benchmarks Index Score 938.9Vultr
BYTE UNIX Benchmarks (Version 5.1.3)
System: vultr.guest: GNU/Linux
OS: GNU/Linux -- 4.9.0-11-amd64 -- #1 SMP Debian 4.9.189-3+deb9u2 (2019-11-11)
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Virtual CPU 523cbcdd6ca4 (4788.9 bogomips)
x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
09:36:01 up 23 min, 1 user, load average: 0.39, 0.18, 0.11; runlevel 5
------------------------------------------------------------------------
Benchmark Run: Sat Jan 18 2020 09:36:01 - 10:04:06
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 25349335.0 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4294.3 MWIPS (10.0 s, 7 samples)
Execl Throughput 3378.6 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 382653.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 107060.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1039637.6 KBps (30.0 s, 2 samples)
Pipe Throughput 561534.3 lps (10.0 s, 7 samples)
Pipe-based Context Switching 89520.1 lps (10.0 s, 7 samples)
Process Creation 7050.7 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5807.6 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 788.3 lpm (60.1 s, 2 samples)
System Call Overhead 329676.0 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 25349335.0 2172.2
Double-Precision Whetstone 55.0 4294.3 780.8
Execl Throughput 43.0 3378.6 785.7
File Copy 1024 bufsize 2000 maxblocks 3960.0 382653.0 966.3
File Copy 256 bufsize 500 maxblocks 1655.0 107060.5 646.9
File Copy 4096 bufsize 8000 maxblocks 5800.0 1039637.6 1792.5
Pipe Throughput 12440.0 561534.3 451.4
Pipe-based Context Switching 4000.0 89520.1 223.8
Process Creation 126.0 7050.7 559.6
Shell Scripts (1 concurrent) 42.4 5807.6 1369.7
Shell Scripts (8 concurrent) 6.0 788.3 1313.8
System Call Overhead 15000.0 329676.0 219.8
========
System Benchmarks Index Score 753.3令人出乎意料的是，pacificrack 的性能竟然好于 Vultr。这对于一些比较吃 CPU 的应用可能会有些帮助。
2、网络性能测试
Pacificrack
国内：
Node Name Upload Speed Download Speed Latency
Speedtest.net 405.11 Mbit/s 565.16 Mbit/s 2.951 ms
Fast.com 0.00 Mbit/s 130.7 Mbit/s -
Beijing 5G CT 2.02 Mbit/s 4.43 Mbit/s -
Nanjing 5G CT 3.07 Mbit/s 6.61 Mbit/s -
Hefei 5G CT 2.29 Mbit/s 6.84 Mbit/s -
Guangzhou 5G CT 6.54 Mbit/s 11.11 Mbit/s -
TianJin 5G CU 29.17 Mbit/s 8.29 Mbit/s -
Shanghai 5G CU 18.94 Mbit/s 21.42 Mbit/s -
Guangzhou 5G CU 58.49 Mbit/s 8.61 Mbit/s -
Tianjin 5G CM 3.56 Mbit/s 107.28 Mbit/s -
Wuxi 5G CM 2.84 Mbit/s 6.72 Mbit/s -
Nanjing 5G CM 0.85 Mbit/s 12.11 Mbit/s -
Hefei 5G CM 14.96 Mbit/s 124.84 Mbit/s -
Changsha 5G CM 3.59 Mbit/s 24.15 Mbit/s -
国外：
Node Name IPv4 address Download Speed
CacheFly 205.234.175.175 31.4MB/s
Linode, Tokyo2, JP 139.162.65.37 9.71MB/s
Linode, Singapore, SG 139.162.23.4 5.74MB/s
Linode, London, UK 176.58.107.39 8.00MB/s
Linode, Frankfurt, DE 139.162.130.8 1.36MB/s
Linode, Fremont, CA 50.116.14.9 17.3MB/s
Softlayer, Dallas, TX 173.192.68.18 1.05MB/s
Softlayer, Seattle, WA 67.228.112.250 981KB/s
Softlayer, Frankfurt, DE 159.122.69.4 386KB/s
Softlayer, Singapore, SG 119.81.28.170 337KB/s
Softlayer, HongKong, CN 119.81.130.170 290KB/s Vultr
国内：
Node Name Upload Speed Download Speed Latency
Speedtest.net 2372.69 Mbit/s 1957.47 Mbit/s 1.263 ms
Fast.com 0.00 Mbit/s 151.3 Mbit/s -
Beijing 5G CT 8.36 Mbit/s 3.99 Mbit/s -
Nanjing 5G CT 67.04 Mbit/s 342.91 Mbit/s -
Hefei 5G CT 10.98 Mbit/s 133.37 Mbit/s -
Guangzhou 5G CT 1.81 Mbit/s 28.90 Mbit/s -
TianJin 5G CU 22.65 Mbit/s 39.31 Mbit/s -
Shanghai 5G CU 111.81 Mbit/s 7.82 Mbit/s -
Guangzhou 5G CU 97.45 Mbit/s 192.28 Mbit/s -
Tianjin 5G CM 0.39 Mbit/s 1.99 Mbit/s -
Wuxi 5G CM 4.21 Mbit/s 5.73 Mbit/s -
Nanjing 5G CM 1.39 Mbit/s 37.00 Mbit/s -
Hefei 5G CM 67.43 Mbit/s 153.11 Mbit/s -
Changsha 5G CM 16.73 Mbit/s 177.33 Mbit/s -
国外：
Node Name IPv4 address Download Speed
CacheFly 205.234.175.175 66.1MB/s
Linode, Tokyo2, JP 139.162.65.37 17.9MB/s
Linode, Singapore, SG 139.162.23.4 13.1MB/s
Linode, London, UK 176.58.107.39 13.9MB/s
Linode, Frankfurt, DE 139.162.130.8 13.2MB/s
Linode, Fremont, CA 50.116.14.9 75.8MB/s
Softlayer, Dallas, TX 173.192.68.18 33.8MB/s
Softlayer, Seattle, WA 67.228.112.250 39.6MB/s
Softlayer, Frankfurt, DE 159.122.69.4 4.88MB/s
Softlayer, Singapore, SG 119.81.28.170 7.42MB/s
Softlayer, HongKong, CN 119.81.130.170 8.41MB/s 显然，pacificrack 的网速实在是过于缓慢，无论是在国内还是国外，都被 Vultr 甩了几条街。尤其是对于移动和电信的用户，更是惨不忍睹。（当时买的时候说是 CN2 VPS，现在看来是一条假 CN2 了。）
还想得知更详细信息的朋友可以访问： pacificrack Vultr，里面有了各种各样的测试结果，包括网速、路由、Ping 等。
三、总结
总的来说，pacificrack 的 VPS 虽然卖的便宜，但各项性能都不太好，特别是具有网络连接缓慢这个致命伤。打算用它来建站的朋友可得小心了，不过联通用户拿来停飞机倒是一个不错的选择。
各位想要建站的朋友可以考虑 Vultr 。新用户用我的 aff 可以获得$100的体验金噢。链接：https://www.vultr.com/?ref=8379290-6G
？？？为什么会冒出烟花
因为本站用的是一位大佬的主题噢，项目具体信息可以在页脚找到。
用vlutr搭建的ssr，电信网测速只有不到1mbps的速度…有点失望