Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

为什么相同连接数的情况 网关负载却不一样? #40

Open
k79e opened this issue Mar 31, 2021 · 3 comments
Open

为什么相同连接数的情况 网关负载却不一样? #40

k79e opened this issue Mar 31, 2021 · 3 comments

Comments

@k79e
Copy link

k79e commented Mar 31, 2021

比如这个跑100个连接 网关rtt有大波动 乃至丢包
ncrack跑200个 rtt基本不咋动

这连接数要是都100 还是一个卡一个不卡
真是奇怪了. 是这个对time_wait 重用支持不好么?

@shadow1ng
Copy link
Owner

我回头研究一下这个,有更详细的测试过程嘛

@shadow1ng
Copy link
Owner

time_wait 这个问题我也没想好怎么解决,已经port.close()之后还是会存在

@k79e
Copy link
Author

k79e commented Mar 31, 2021

有 我用的ip列表 几千个数量 然后字典用的500条的
-t 测试用的话必须>100 200-400比较好

fscan -m ssh -np -p 22 -user root -pwdf /sft/dict/dict.txt -hf /data/list -time 3 -t 100
然后我写了个脚本 能监控wait / time_wait数量
可是2个程序数量都差不多 这让人反而更疑惑了

这是ncrack的
SYN=20 EST=237 WAIT=2322 TW=2276 FIN=27
SYN=18 EST=234 WAIT=2262 TW=2215 FIN=26
SYN=29 EST=229 WAIT=2315 TW=2267 FIN=28
SYN=20 EST=235 WAIT=2308 TW=2263 FIN=25

From 192.168.1.1: bytes=60 seq=0001 TTL=64 ID=c5d5 time=9.541ms
From 192.168.1.1: bytes=60 seq=0002 TTL=64 ID=c5d6 time=15.769ms
From 192.168.1.1: bytes=60 seq=0003 TTL=64 ID=c5d7 time=1.717ms
From 192.168.1.1: bytes=60 seq=0004 TTL=64 ID=c5d8 time=1.326ms
From 192.168.1.1: bytes=60 seq=0005 TTL=64 ID=c5d9 time=1.283ms
From 192.168.1.1: bytes=60 seq=0006 TTL=64 ID=c5da time=0.981ms
From 192.168.1.1: bytes=60 seq=0007 TTL=64 ID=c5db time=0.820ms
From 192.168.1.1: bytes=60 seq=0008 TTL=64 ID=c5dc time=58.704ms
From 192.168.1.1: bytes=60 seq=0009 TTL=64 ID=c5dd time=1.288ms
From 192.168.1.1: bytes=60 seq=000a TTL=64 ID=c5de time=0.992ms
rtt就小有波动然后立刻恢复了

fscan的我等会补下数据
连续跑了一阵 现在是
SYN=29 EST=226 WAIT=2288 TW=2205 FIN=66
SYN=17 EST=237 WAIT=2251 TW=2168 FIN=62
SYN=14 EST=241 WAIT=2285 TW=2202 FIN=63
SYN=13 EST=242 WAIT=2223 TW=2140 FIN=62

fscan
SYN=29 EST=73 WAIT=2205 TW=2192 FIN=16
SYN=26 EST=78 WAIT=2258 TW=2251 FIN=8
SYN=23 EST=82 WAIT=2310 TW=2304 FIN=7
SYN=31 EST=71 WAIT=2374 TW=2360 FIN=15

网关延时目前没什么异常 (但是连接很小 time_wait却和ncrack跑200多并发时候一样)
SYN=32 EST=171 WAIT=99 TW=8 FIN=4
SYN=24 EST=178 WAIT=224 TW=113 FIN=24
SYN=60 EST=143 WAIT=370 TW=220 FIN=61
SYN=40 EST=160 WAIT=494 TW=362 FIN=49
SYN=38 EST=166 WAIT=620 TW=499 FIN=40
SYN=43 EST=159 WAIT=754 TW=624 FIN=42
SYN=48 EST=157 WAIT=885 TW=759 FIN=40
SYN=49 EST=164 WAIT=1009 TW=888 FIN=34

上面是重新开-t 200的参数 网关立刻卡了
From 192.168.1.1: bytes=60 seq=043f TTL=64 ID=ca13 time=44.681ms
From 192.168.1.1: bytes=60 seq=0440 TTL=64 ID=ca14 time=79.018ms
From 192.168.1.1: bytes=60 seq=0441 TTL=64 ID=ca15 time=77.313ms
From 192.168.1.1: bytes=60 seq=0442 TTL=64 ID=ca16 time=55.437ms
From 192.168.1.1: bytes=60 seq=0443 TTL=64 ID=ca17 time=84.720ms
From 192.168.1.1: bytes=60 seq=0444 TTL=64 ID=ca18 time=94.563ms
From 192.168.1.1: bytes=60 seq=0445 TTL=64 ID=ca19 time=71.428ms
From 192.168.1.1: bytes=60 seq=0446 TTL=64 ID=ca1a time=102.685ms
From 192.168.1.1: bytes=60 seq=0447 TTL=64 ID=ca1b time=97.516ms
From 192.168.1.1: bytes=60 seq=0448 TTL=64 ID=ca1c time=101.325ms
From 192.168.1.1: bytes=60 seq=0449 TTL=64 ID=ca1d time=125.576ms
From 192.168.1.1: bytes=60 seq=044a TTL=64 ID=ca1e time=87.738ms
From 192.168.1.1: bytes=60 seq=044b TTL=64 ID=ca1f time=104.736ms
From 192.168.1.1: bytes=60 seq=044c TTL=64 ID=ca20 time=139.213ms
From 192.168.1.1: bytes=60 seq=044d TTL=64 ID=ca21 time=126.899ms
From 192.168.1.1: bytes=60 seq=044e TTL=64 ID=ca22 time=103.964ms

重新开了下 这时候就卡了 网关rtt连续变成好几十
SYN=66 EST=138 WAIT=125 TW=34 FIN=67
SYN=32 EST=169 WAIT=238 TW=180 FIN=39
SYN=41 EST=159 WAIT=398 TW=330 FIN=47
SYN=45 EST=153 WAIT=543 TW=483 FIN=37
SYN=34 EST=165 WAIT=676 TW=616 FIN=37
很奇怪 time_wait当时不怎么高 而且这个TW本身就是要卡一阵才消失
网络变卡的时候 关了程序 那些TW还在 但是网关负载顿时下去了.
这总不会是数据传输造成的卡顿把. 让人想不明白. 可是其他程序也是用这个并发 比他还大 反而没问题.
让我感觉莫名其妙的.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants