Hi Saul,
On Mon, Mar 11, 2013 at 05:29:10PM -0400, Saul Waizer wrote:
> Absolutley, here is the config:
>
> global
> log 127.0.0.1 local0 crit
> pidfile /var/run/haproxy.pid
> stats socket /var/run/haproxy.stat mode 666
> nbproc 2
> maxconn 65000
> tune.maxrewrite 1024
> tune.bufsize 32768
> user haproxy
> group haproxy
> daemon
>
>
> defaults
> mode http
> log global
> option dontlognull
> option httplog
> option http-server-close
> #option forwardfor except 127.0.0.0/8
> option forwardfor header X-Real-IP
> option redispatch
> stats enable
> stats auth change:me
> stats uri /stav1
> timeout connect 5000 # default 5 second time out if a backend is not
> found
> timeout client 300s
> timeout server 300s
> #timeout http-request 10s
> #timeout queue 1m
> #timeout http-keep-alive 10s
> timeout check 5s
> maxconn 65000
> retries 3
>
> frontend sinatra *:5000
> default_backend sinatra_backend
>
> backend sinatra_backend
> mode http
> balance roundrobin
> option httpchk HEAD /ha.txt HTTP/1.0
Could you please run "strace -tt -p $PID" while it's doing
this ?
Also, I suspect you're having a few servers in the backend.
Could you please try to :
- comment out the "timeout check"
- comment out "nbproc 2"
Neither of these should cause any issue, but you never know.
Also, a number of bugs were fixed since 1.4.22 in the latest snapshot,
waiting for more important ones to do a release (though I think I could
do one anyway). One of them concerns the pollers where an event might
remain present. I diagnosed that it should not cause such issues but it
is possible that I have overlooked it. Could you then test this latest
snapshot from there to verify if you get any difference :
http://haproxy.1wt.eu/download/1.4/src/snapshot/
Thanks!
Willy
