haproxy Infinite page refresh loop - haproxy

Question

On different hosts, there are two identical sites on IIS web servers, access to them is open on two ports 443 and 9000, if I turn on both hosts in haproxy, then when the site page is opened, an endless page refresh cycle occurs. It works if you disable one of the catches or specify the weight as srv1 weight 100 srv2 weight 0. How can I fix this?
My haproxy config:
global
# to have these messages end up in /var/log/haproxy.log you will
# need to:
#
# 1) configure syslog to accept network log events. This is done
# by adding the '-r' option to the SYSLOGD_OPTIONS in
# /etc/sysconfig/syslog
#
# 2) configure local2 events to go to the /var/log/haproxy.log
# file. A line like the following can be added to
# /etc/sysconfig/syslog
#
# local2.* /var/log/haproxy.log
#
log 127.0.0.1 local2
# log /dev/log local0
# log /dev/log local1 notice
chroot /var/lib/haproxy
pidfile /var/run/haproxy.pid
maxconn 10000
user haproxy
group haproxy
daemon
# turn on stats unix socket
stats socket /var/lib/haproxy/stats mode 666 level user
ssl-default-bind-options no-sslv3
ssl-default-bind-ciphers EECDH:+AES256:-3DES:RSA+AES:RSA+3DES:!NULL:!RC4
# ssl-server-verify none
defaults
mode http
log global
option httplog
option dontlognull
#option http-server-close
#option forwardfor except 127.0.0.0/8
#option redispatch
retries 10
timeout http-request 10s
timeout queue 1m
timeout connect 10s
timeout client 1m
timeout server 150s
timeout http-keep-alive 65s
timeout check 10s
maxconn 10000
log-format "%{+Q}o\ client = %ci:%cp, server = %si:%sp path = %HU, status = %ST, %b, t_простоя = %Ti, t_отклика = %Tr, t_сеанса = %Tt, request = %r, byte = %B"
frontend stats
bind *:8404
http-request use-service prometheus-exporter if { path /metrics }
no log
stats enable
stats uri /stats
stats refresh 10s
stats auth adm:123
frontend http
bind *:80
mode http
redirect scheme https if !{ ssl_fc }
frontend https
bind *:443 ssl crt /etc/haproxy/tls/haproxy.pem alpn h2,http/1.1
bind *:9000 ssl crt /etc/haproxy/tls/haproxy.pem alpn h2,http/1.1
option forwardfor
option http-keep-alive
http-request add-header X-Forwarded-Proto https
http-request set-header X-Client-IP %[src]
errorloc 404 https://xxxx.com/500
errorloc 500 https://xxxx.com/500
errorloc 503 https://xxxx.com/500
errorloc 504 https://xxxx.com/500
use_backend https if { hdr(host) -i xxxx.com }
use_backend 9000 if { hdr(host) -i xxxx.com:9000 }
backend https
balance roundrobin
server srv1 192.168.1.11:443 check cookie srv1 weight 80 ssl verify none
server srv2 192.168.1.111:443 check cookie srv2 weight 20 ssl verify none
backend 9000
balance roundrobin
server srv1 192.168.1.11:9000 check cookie srv1 weight 80 ssl verify none
server srv2 192.168.1.111:9000 check cookie srv2 weight 20 ssl verify none

Answer 1

You are using the "cookie" strategy for loadbalancing, but no way to set this cookie :
balance roundrobin
server srv1 192.168.1.11:443 check cookie srv1 weight 80 ssl verify none
server srv2 192.168.1.111:443 check cookie srv2 weight 20 ssl verify none
As a result, the client communicates with both server alternatively. This does not suit well with IIS (and/or underlying technos).
Simply tell HAProxy which cookie to use/set. For example, a cookie named SERVERID :
balance roundrobin
cookie SERVERID insert indirect nocache
server srv1 192.168.1.11:443 check cookie srv1 weight 80 ssl verify none
server srv2 192.168.1.111:443 check cookie srv2 weight 20 ssl verify none
If the client already has a SERVERID cookie, then HAProxy sends the traffic to the corresponding server.
If the client does not have a SERVERID cookie, then HAProxy chooses and instructs the client to Set-Cookie: SERVERID=s1 for example.
Source : https://www.haproxy.com/blog/load-balancing-affinity-persistence-sticky-sessions-what-you-need-to-know/