# help to configure a double failover with 3 node cluster

**URL:** <https://forums.percona.com/t/help-to-configure-a-double-failover-with-3-node-cluster/2995>\
**Category:** Percona XtraDB Cluster 5.x\
**Created:** [October 9, 2013, 6:44am UTC](https://forums.percona.com/t/help-to-configure-a-double-failover-with-3-node-cluster/2995 "2013-10-09T06:44:48Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![bjornlieb](https://avatars.discourse-cdn.com/v4/letter/b/ea5d25/32.png) [@bjornlieb](https://forums.percona.com/u/bjornlieb)\
**Post date:** [October 9, 2013, 6:44am UTC](https://forums.percona.com/t/help-to-configure-a-double-failover-with-3-node-cluster/2995/1 "2013-10-09T06:44:48Z")

</div>

Hi, we are looking to configure a double failover scheme using a 3 node cluster.

Node 1, Node 2, Node 3

So node 1 will take all the traffic in normal conditions,  
if node 1 fails then node 2 will take all the traffic,  
if node 2 fails then node 3 will take all the traffic.

So far I can implement the first two requirements using the “Backup” option for nodes 2 and 3.

_but_ I have found that if node 1 and 2 fails, then node 3 does not take over the traffic, and the server connection is lost. Is there any reason why node 3 does not operate?

(the HAProxy stats web page does show correctly the three nodes with 2 and 3 as backup)

* * *

below is our configuration file

global  
stats socket /var/run/haproxy.sock mode 0600 level admin

log 127.0.0.1 local2 debug  
#chroot /var/lib/haproxy  
pidfile /var/run/haproxy.pid  
maxconn 4000  
daemon

defaults  
mode tcp  
log global  
option tcplog  
timeout connect 10000 # default 10 second time out if a backend is not found  
timeout client 300000  
timeout server 300000  
maxconn 20000

# For Admin GUI

listen stats  
bind :8080  
mode http  
stats enable  
stats uri /stats

listen mysql \*:3306  
mode tcp  
balance roundrobin  
option mysql-check user haproxyUser  
option log-health-checks  
server MySQL-NodeA 10. **.**.**:3306 check  
server MySQL-NodeB 10. **.**.**:3306 check backup  
server MySQL-NodeC 10. **.**.\*\*:3306 check backup

---

<div class="post-metadata">

**Author:** ![przemek](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/przemek/32/3_2.png) [@przemek](https://forums.percona.com/u/przemek)\
**Post date:** [October 26, 2013, 4:33am UTC](https://forums.percona.com/t/help-to-configure-a-double-failover-with-3-node-cluster/2995/2 "2013-10-26T04:33:41Z")

</div>

If the node2 and node3 will fail in pretty much the same time, then the node1 will detect the situation as potential split brain and will go into non-Primary state, hence no queries will be accepted. To solve this, you should configure also an arbitrator node, maybe on the same host as HAproxy.