1. 'unavailable' means the server will automatically and temporarily be
removed from the pool of servers managed by the upstream (same effect as
manually parametering it as down)
2. the fail_timeout parameter documentation of the very section you
provided a link to is pretty clear: this value both configure the duration
during which the parametered amount of failure must be recorded and the
time during which the server will subsequently be unavailable

---
*B. R.*

On Wed, Jun 21, 2017 at 3:54 PM, linbo liao <llbgurs@gmail.com> wrote:

> Hi,
>
> Refer to http://nginx.org/en/docs/http/ngx_http_upstream_module.html#server
> , if all primary server are unavailable, backup server will handle request.
>
> I have two question?
> 1. What's the meaning of unavailable?
> 2. When will primary server come back, after fail_timeout?
>
> Thanks,
> Linbo
>
>
>
> _______________________________________________
> nginx mailing list
> nginx@nginx.org
> http://mailman.nginx.org/mailman/listinfo/nginx
>
_______________________________________________
nginx mailing list
nginx@nginx.org
http://mailman.nginx.org/mailman/listinfo/nginx