That's because HAproxy does not detect instantly that a backend server is down. You have to set the intervals.
On Tue, Jan 22, 2013 at 11:59 AM, Boris Epstein <[email protected]>wrote: > Hello all, > > I just configured a very primitive HAProxy installation on a CentOS 6 > machine with the configuration that looks as follows: > > > ---------------------------------------------------------------------------------------------------------------------------------------- > # Simple configuration for an HTTP proxy listening on port 80 on all > # interfaces and forwarding requests to a single backend "servers" with a > # single server "server1" listening on 127.0.0.1:8000 > global > daemon > maxconn 1024 > > defaults > mode http > timeout connect 5000ms > timeout client 50000ms > timeout server 20000ms > > frontend http-in > bind *:80 > option http-server-close > default_backend servers > > backend servers > balance roundrobin > server server1 10.12.204.18 check port 80 maxconn 32 > server server1 10.12.204.19 check port 80 maxconn 32 > > > ---------------------------------------------------------------------------------------------------------------------------------------- > > Everything seems to be OK but one pesky little problem: every time you > shut off the HTTP server on either 10.12.204.18 or 10.12.204.19 and reload > the HAProxy's URL on the external IP it skips a beat - i.e., you get the > 503 error. But then you reload again and the remaining server kicks into > action. > > Has anybody seen that? Does anybody have any idea as to why this would be? > > Thanks. > > Boris. > -- zachary alex stern I systems architect o: 212.363.1654 x106 | f: 212.202.6488 | [email protected] 60-62 e. 11th street, 4th floor | new york, ny | 10003 www.enternewmedia.com
