ubuntu 12.04 : ipvsadm , keepalived

Here's an explanation of how to configure ipvsadm and keepalived.

client

 |

 | VIP 192.168.10.112 , real ip : 192.168.10.111

LVS

 | real ip : 192.168.30.111

 |

Web*2 ( 192.168.30.10 , 192.168.30.20 )

# tail -1 /etc/lsb-release DISTRIB_DESCRIPTION="Ubuntu 12.04.2 LTS" # uname -ri 3.2.0-39-generic x86_64

install

# apt-get install -y ipvsadm keepalived

start ipvsadm.

at this time , there are no Virtual seervers and nodes.

# ipvsadm -ln IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn

configure keepalived.conf ( /etc/keepalived/keepalived.conf )

# cat /etc/keepalived/keepalived.conf virtual_server_group HTTP 100 {        192.168.10.112 80 } virtual_server group HTTP 100 {        delay_loop 3        lvs_sched rr        lvs_method NAT        protocol TCP        virtualhost foo.bar        sory_server 192.168.30.30 80        real_server 192.168.30.10 80 {        weight 1        inhibit_on_failure        HTTP_GET {                url {                        path /index.html                        status_code 200                }                connect_timeout 3        } }        real_server 192.168.30.20 80 {        weight 1        inhibit_on_failure        HTTP_GET {                url {                        path /index.html                        status_code 200                }                connect_timeout 3        } } }

start keepalived

# /etc/init.d/keepalived start * Starting keepalived keepalived                                        [ OK ]

the virtual server is added

# ipvsadm -ln IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn TCP  192.168.10.112:80 rr  -> 192.168.30.10:80             Masq    1      0          0  -> 192.168.30.20:80             Masq    1      0          0

# curl http://192.168.10.112 centos6-3 # curl http://192.168.10.112 centos6-4

cap data on the web server. the following is heal check packets from LVS to Real server.

192.168.30.111 : LVS’s IP , 192.168.30.10 : Real Server’s IP

# tshark -i eth0 port 80 -R '(ip.addr==192.168.30.111)' | grep -i get Capturing on eth0  0.000546 192.168.30.111 -> 192.168.30.10 HTTP GET /index.html HTTP/1.0  3.002883 192.168.30.111 -> 192.168.30.10 HTTP GET /index.html HTTP/1.0  6.005483 192.168.30.111 -> 192.168.30.10 HTTP GET /index.html HTTP/1.0  9.008175 192.168.30.111 -> 192.168.30.10 HTTP GET /index.html HTTP/1.0

when the apache goes down.

/var/log/syslog ( LVS )

Apr  2 20:23:38 ubuntu-lvs Keepalived_healthcheckers: Error connecting server [192.168.30.10]:80. Apr  2 20:23:38 ubuntu-lvs Keepalived_healthcheckers: Disabling service [192.168.30.10]:80 from VS [HTTP]:0

when the apache goes up

/var/log/syslog ( LVS )

Apr  2 20:26:32 ubuntu-lvs Keepalived_healthcheckers: HTTP status code success to [192.168.30.10]:80 url(1). Apr  2 20:26:35 ubuntu-lvs Keepalived_healthcheckers: Remote Web server [192.168.30.10]:80 succeed on service. Apr  2 20:26:35 ubuntu-lvs Keepalived_healthcheckers: Enabling service [192.168.30.10]:80 to VS [HTTP]:0