虚拟节点会有问题吗？

[superen7527]

抱歉 不知道在哪里问问题 好像只能这里了。要是不符合规定就关闭issue吧。

上一个issue 试了一下虚拟节点，可以用，但是有两个问题：

1. 随机random 很大概率同一个ip连续出口2+3次 当然可能是我用while true判断，循环的太快了，但是不是说明随机的这个值缓存的地方还是怎么实现会导致短时间太快请求会复用，还是说只是刚好随机的巧合？（如果是复用，round可能也会？我还没试过 感觉round没有随机实现的复杂，每次请求后应该百分百会变，我后面试试再来更新issue）

2. 通过这个虚拟节点的池访问外部，会概率出现超时的情况。我测试过：不用auto指定监听协议指定http/socks，v4/v6，更换目标请求地址，都会有这个问题。( http好像就单纯卡住前台都没报错，socks才会返回 Can't complete SOCKS5 connection 这个万能报错。)
   详细日志 （说实话搞了半天才定位到这里 日志不太好找，而且没记录name和出口ip）

{"handler":"socks5","kind":"handler","level":"error","listener":"tcp","msg":"route(retry=0) dial tcp [fe80::248:a7ff:fe3e:426c]:0->[2a09:0:13::9]:80: i/o timeout","service":"v6","time":"2024-07-08T03:46:14.760+08:00"}
{"duration":15015287763,"handler":"socks5","kind":"handler","level":"info","listener":"tcp","local":"127.0.0.1:666","msg":"127.0.0.1:43196 >< 127.0.0.1:666","remote":"127.0.0.1:43196","service":"v6","time":"2024-07-08T03:46:14.760+08:00"}
{"handler":"socks5","kind":"service","level":"error","listener":"tcp","msg":"dial tcp [fe80::248:a7ff:fe3e:426c]:0->[2a09:0:13::9]:80: i/o timeout","service":"v6","time":"2024-07-08T03:46:14.760+08:00"}
{"handler":"socks5","kind":"handler","level":"info","listener":"tcp","local":"127.0.0.1:666","msg":"127.0.0.1:42996 <> 127.0.0.1:666","remote":"127.0.0.1:42996","service":"v6","time":"2024-07-08T03:46:14.766+08:00"}
{"cmd":"connect","dst":"[2a09:0:13::9]:80/tcp","handler":"socks5","kind":"handler","level":"info","listener":"tcp","local":"127.0.0.1:666","msg":"127.0.0.1:42996 <-> [2a09:0:13::9]:80","remote":"127.0.0.1:42996","service":"v6","time":"2024-07-08T03:46:14.924+08:00"}

---

现在就是不知道是不是虚拟节点的问题，还是gost的问题，还是我服务器的问题。issue没找到相同的，，不知道是不是bug，稍微反馈一下。
或者能否优化一下，如果遇到这种超时，马上换一个出口重新访问？我看了很久文档 只有一个failTimeout超时禁用以后出口的api，而且还是默认开启，没有这种超时知道换出口的逻辑。