Replies: 30 comments 20 replies
-
可以提供一下配置文件(删除敏感信息),以及具体的服务器的位置么? |
Beta Was this translation helpful? Give feedback.
-
@xiaokangwang 这是我的一些信息供参考 如果有问题还请指出
理论上应该这样服务端就ok了? |
Beta Was this translation helpful? Give feedback.
-
我也碰到了,改了传输协议未kcp就莫名好了。inbound 下面加 怀疑是GFW能检测到 tcp的vmess协议,直接把包给drop了。 |
Beta Was this translation helpful? Give feedback.
-
能说的更详细一点吗?为什么传输层协议改成kcp居然没事,还有这是客户端还是服务器端的配置文件。你说的方法我还没有尝试,但看起来并不是很靠谱 |
Beta Was this translation helpful? Give feedback.
-
都要改,可以看这里。https://www.v2fly.org/config/transport/mkcp.html#kcpobject |
Beta Was this translation helpful? Give feedback.
-
出问题的貌似都是vultr的服务器,具体表现就是ip正常,端口正常,所有的都正常,但网络不通。测试延迟会显示context deadline exceeded。换端口无用。应该是所有使用v2ray tcp模式的包全被精准识别然后丢掉了。 配置用的是tcp |
Beta Was this translation helpful? Give feedback.
-
这很难下定论,首先 Vultr 的服务器与大陆之间的路由本身并不友好,以及 Vultr 因为购买门槛低等原因导致其 IP 段很可能被 GFW 重点关注。如果 GFW 检测到无法识别的协议从大陆发送到 Vultr 机房,被丢包也并不是不可能。 需要一些试验数据来验证这些猜想是否成立。 |
Beta Was this translation helpful? Give feedback.
-
怀疑是vultr被盯上了,所有相关软件版本都是最新 |
Beta Was this translation helpful? Give feedback.
-
我是买 的阿里云新加坡的 轻量服务也不行 |
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
-
确实我一开始也觉得是自己的问题,后来一想不对劲。我的服务器用了1年半了,自从使用v2ray的tcp协议搭配udp2raw后就从来没出过问题,结果最近突然连接很困难。一检查发现不仅ip没被封,连端口都很正常。 然后发现很多人跟我一样的情况,基本就能确定是v2ray的tcp协议被识破了。目前几个issue里面提到的类似问题,有好几个是用的vultr,并且只有tcp协议会出问题,更换为mkcp后就正常了。 |
Beta Was this translation helpful? Give feedback.
-
可能导致连接被阻断的因素有很多,目前的信息还不足以证明 Vmess 能被识别。因为 mKCP 基于 UDP 协议,所以 GFW 对 TCP 和 UDP 的处理规则可能有所不同。 如果实在不放心 Vmess 协议的 TCP 传输模式,建议尝试为其启用 AEAD 和 TLS。 |
Beta Was this translation helpful? Give feedback.
-
遇到同样的问题,kcp 也没效 |
Beta Was this translation helpful? Give feedback.
-
有解决了的兄弟顶一下哈 |
Beta Was this translation helpful? Give feedback.
-
一台vultr的vps,前几天发生上面问题,测试vmess+tcp传输几乎秒封端口(nc测试time out),而且只封端口,不封ip,多久解封不清楚,反正再连再秒封。改用vmess+tcp+tls可以正常稳定连接,另外测试对于vless协议不存在这个问题,只用vless+tcp也不会有任何问题,可以正常稳定连接。所以个人认为可以确认vmess明码可以被精确定位打击。 |
Beta Was this translation helpful? Give feedback.
-
昨天我尝试重现这个对于VMess和其他随机协议的封锁,但是无法重现这个问题。 我在Vultr日本,Digital Ocean SFO3(旧金山) 搭建了4种服务器: VMess+TCP, Shadowsockets, 随机数据流, HTTP, 并在超过半个小时的时间内访问这些服务器。然而并没有检测到对这些服务的封锁。 在Vultr日本服务器上运行的Shadowsockets和随机数据流受到了重放探测,探测的方式类似于在GFWReport报告中的R1(原样重放), 以及R2,R3(修改后重放)。估计在GFWReport给出详细报告后可以得到更详细的重现封锁的方法。对于无法重现这个封锁的事情有3种可能的解释:不同的地区,运营商,用户的限制级别不同;不同的时间限制的级别不同(如开会时封锁会更严格);以及流量的形状,即在测试时产生的测试流量不会触发针对特性套嵌流量的封锁。 Yesterday, I tried to reproduce that random stream look like nothing proxy blocking behaviour, but it was unsuccessful. All 4 types of service: VMess+TCP, Shadowsockets, Random Stream, HTTP setup at Vultr Japan, Digital Ocean SFO3 is accessible after at least half an hour of usage. Active probe([Identical replay=R1], Partical Replay=R2, R3) are observed at Shadowsockets and Random Stream Service. More information about how to reproduce this may become available once the full report from gfw_report is released. There are 3 possible hypotheses for the non-reproducible characteristic: Different regions and network users might have a different attention score(Vary: User-Identity); This blocking policy is only enacted during sensitive periods like the Party's meeting(Vary: Time); The exact type of traffic determines the blocking behaviour, the traffic synthetic traffic during experiment do not trigger the blocking of its wrapping traffic(Vary: Traffic Shape). |
Beta Was this translation helpful? Give feedback.
-
这次是11月初对某些VPS提供商的集体端口干扰 应该是不太好复现的 |
Beta Was this translation helpful? Give feedback.
-
重点机房被干扰,很多人用不同的连接方式都被观察到了干扰。 |
Beta Was this translation helpful? Give feedback.
-
V2Ray-noTLS和SS都沒有金鑰交換。 |
Beta Was this translation helpful? Give feedback.
-
对11 月开会后 一直延续到现在。这里一台受到 间歇性 tcp 阻断的是阿里云 HK 轻量 47.xxxxxx段。 |
Beta Was this translation helpful? Give feedback.
-
vless+tls+tcp稳定得很,这配置从升级vless到现在,没阻断过。 |
Beta Was this translation helpful? Give feedback.
-
我的vmess+ws+tls在近期突然就访问不通了,一样的现象客户端报错,改了AEAD也没用,但是同一台服务器上的vless+tls+tcp就不受影响。很迷,找不到原因。 |
Beta Was this translation helpful? Give feedback.
-
最近搞了两个轻量应用,第一个安装的时候是4.43似乎还正常,前两天安装版本是4.44的时候,无论如何都打不开了 |
Beta Was this translation helpful? Give feedback.
-
稳定用了一年多的配置最近也不行了 IP+端口 telnet都能通,但v2ray客户端访问就是连不上 app/proxyman/outbound: failed to process outbound traffic > proxy/vmess/outbound: failed to find an available destination > common/retry: [dial tcp -.-.-.65:443: i/o timeout dial tcp -.-.-.65:443: operation was canceled] > common/retry: all retry attempts failed |
Beta Was this translation helpful? Give feedback.
-
谢谢!---- 回复的原邮件 ***@***.***>发送日期2022年02月14日 20:24 ***@***.***> ***@***.******@***.***>主题Re: [v2fly/v2ray-core] 2021年11月份开始V2Ray很难成功访问成功 (Discussion #1383)
Jrohy大神的multi-v2ray脚本:source <(curl -sL https://multi.netlify.app/v2ray.sh) --zh
—Reply to this email directly, view it on GitHub, or unsubscribe.Triage notifications on the go with GitHub Mobile for iOS or Android.
You are receiving this because you commented.Message ID: ***@***.***>
[
{
***@***.***": "http://schema.org",
***@***.***": "EmailMessage",
"potentialAction": {
***@***.***": "ViewAction",
"target": "#1383 (reply in thread)",
"url": "#1383 (reply in thread)",
"name": "View Discussion"
},
"description": "View this Discussion on GitHub",
"publisher": {
***@***.***": "Organization",
"name": "GitHub",
"url": "https://github.com"
}
}
]
|
Beta Was this translation helpful? Give feedback.
-
我的vmss+tcp运行正常。 |
Beta Was this translation helpful? Give feedback.
-
一般新机器刚搭完就挂的,一般是 IP 有问题,看看能不能直接换一个 IP |
Beta Was this translation helpful? Give feedback.
-
我也用的是233的脚本,把协议改成kcp就好了 |
Beta Was this translation helpful? Give feedback.
-
最新消息... 人在法国,搭建反向代理的时候被反代的在法国(oracle),做反代的在日本(vultr),他们之间打tunnel的时候,怎么试都不成功,上面说的mkcp、别的比如http、ws等等都试过了,不行...... 最后换成websocket+tls+web可以了 所以结论就是,**vultr拦截了别的这些协议,就https没办法拦截而已 |
Beta Was this translation helpful? Give feedback.
-
偶尔能打开一个页面,很慢很慢,大家怎么解决的
Beta Was this translation helpful? Give feedback.
All reactions