Re: option redir and latest snapshot

Willy Tarreau a écrit :
> On Fri, Feb 15, 2008 at 10:01:47AM +0100, florian iragne wrote:

>> Willy Tarreau a écrit :
>>>> no, less than 10s. e.g., check inter 10000 fall 4 rise 1 works
>>>> while check inter 20000 fall 4 rise 1 gives the strange behaviour i 
>>>> mention.
>> Well, i've made a test, and now, it's really weird!

> Some timeouts are missing and it looks like haproxy considers them as zero.
> I had to set both "timeout connect" and "timeout checks". If either is missing,
> the check fails. The fact that there is an interaction between your backends
> is just a time race.
>
> I'm currently fixing this enormous bug, but in the mean time, I suggest adding
> the following in your defaults section :
>
> timeout connect 5s
> timeout check 5s

global

         log 127.0.0.1   local0 notice
         maxconn 32000
	ulimit-n 65536
         user haproxy
         group haproxy

defaults
         log     global
         mode    http
         option  httplog
         option  dontlognull
         retries 3
         option redispatch
         contimeout      5000
         clitimeout      50000
         srvtimeout      50000

listen webfarm-1 :80
         maxconn 32000
         log global
         stats enable
         stats auth toto:toto
         balance
         option httpclose
         option forwardfor
         option httpchk
         option abortonclose
         option httplog
         option logasap

> Also, you have a syntax oddness here :
>

>>         option httpchk HEAD /alive HTTP/1.1\r\nHost: www

oups, i change it right now!

adding the timeout connect/check seems to solve the problem

thanks

Florian Received on 2008/02/15 10:45