a bug,if at the beginning have a down server

[yaoweibin]

a bug,if at the beginning have a down server

[yaoweibin]

This may be not a bug.

You can specify the default_down to false. Then all the servers will be
treated to be UP at the beginning.

I suggest you specify a larger timeout value. 500 miliseconds is too
short, even in the intranet. There may be timeout very often. And the
interval should be larger.

On 2012/5/11 13:28, liutaihua wrote:

eg:
upstream {
server a;
server b; #this is a down server
check interval=1000 rise=3 fall=3 timeout=500 type=tcp default_down=true;
}

if interval time large enough that when a request have arrived, the first request will receive a http 502 response.

sometimes, i don't want a short time interval

---

Reply to this email directly or view it on GitHub:
#11

[liutaihua]

OK, i got it.
thanks you reply.
about the timeout value , i see,i use this module at my internal network