# HAProxy / node 3 doesn't respond

**URL:** <https://forums.percona.com/t/haproxy-node-3-doesnt-respond/2907>\
**Category:** Percona XtraDB Cluster 5.x\
**Created:** [August 26, 2013, 12:52am UTC](https://forums.percona.com/t/haproxy-node-3-doesnt-respond/2907 "2013-08-26T00:52:32Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![liliputien](https://avatars.discourse-cdn.com/v4/letter/l/e9c0ed/32.png) [@liliputien](https://forums.percona.com/u/liliputien)\
**Post date:** [August 26, 2013, 12:52am UTC](https://forums.percona.com/t/haproxy-node-3-doesnt-respond/2907/1 "2013-08-26T00:52:32Z")

</div>

Hi every body,

I try to do xtradbcluster whith HA Proxy. I have 3 node for the cluster. When I cut service mysql on third node, I d’ont have any response by haproxy.

This is the configuration of haproxy :

listen MySQL 0.0.0.0:3306  
timeout connect 10s  
timeout client 1m  
timeout server 1m  
balance roundrobin  
mode tcp  
option mysql-check user haproxy  
server node1 172.16.66.61:3306 check  
server node2 172.16.66.62:3306 check  
server node3 172.16.66.63:3306 check

And the configuration of node1 for exemple :

[mysqld]

datadir=/var/lib/mysql  
user=mysql

auto\_increment\_offset = 4  
auto\_increment\_increment = 4

# Path to Galera library

wsrep\_provider=/usr/lib64/libgalera\_smm.so

# Cluster connection URL contains the IPs of node#1, node#2 and node#3

wsrep\_cluster\_address=gcomm://172.16.66.61,172.16.66.62,172.16.66.63

# In order for Galera to work correctly binlog format should be ROW

binlog\_format=ROW

# MyISAM storage engine has only experimental support

default\_storage\_engine=InnoDB

# This is a recommended tuning variable for performance

innodb\_locks\_unsafe\_for\_binlog=1

# This changes how InnoDB autoincrement locks are managed and is a requirement for Galera

innodb\_autoinc\_lock\_mode=2

# Node #1 address

wsrep\_node\_address=172.16.66.61

# SST method

wsrep\_sst\_method=xtrabackup

# Cluster name

wsrep\_cluster\_name=my\_centos\_cluster

# Authentication for SST method

wsrep\_sst\_auth=“sstuser:s3cret”

# Customisation

tmp\_table\_size = 32M  
max\_heap\_table\_size = 32M  
slow\_query\_log=1  
long\_query\_time=1

query\_cache\_limit = 2M  
query\_cache\_size = 32M

Have you got any idea to know what happen ? or any help to know what happen ?

All node id up and I can red/write data on any node. Replication is ok between all node. Mysql client connection on all node is ok.

But mysql client connection on haproxy when node 1 and node 2 mysql services is down, I have this responses : ERROR 2013 (HY000): Lost connection to MySQL server at ‘reading initial communication packet’, system error: 0

Thanks a lot for your help …

ps : I knows My network is a Satan’s network 172 .16.66.0/24

---

<div class="post-metadata">

**Author:** ![liliputien](https://avatars.discourse-cdn.com/v4/letter/l/e9c0ed/32.png) [@liliputien](https://forums.percona.com/u/liliputien)\
**Post date:** [August 26, 2013, 11:19pm UTC](https://forums.percona.com/t/haproxy-node-3-doesnt-respond/2907/2 "2013-08-26T23:19:07Z")

</div>

I solve the problem : delete server node 3 and create new node.