flyinweb's blog

As my last post, I had built a more healthy load balancer. But it is still not practical. Because if the linux director fails, the system will not work. The high availability should be considered. In this post, I list my steps to implement HA function.

I added another machine to the system. The environment of my lab is as below. 

ha1: eth0: 192.168.26.201 eth1: 192.168.26.221(heartbeat)
ha2: eth0: 192.168.26.202 eth1: 192.168.26.222(heartbeat)
rs1: eth0: 192.168.26.211
rs2: eth0: 192.168.26.212                                              
VIP: 192.168.26.200
gateway: 192.168.26.1 

Configuration
On the ha1 system
Delete the following lines from /etc/rc.local.
ifconfig eth0:0 192.168.26.200 netmask 255.255.255.255 broadcast 192.168.26.200 up
echo 1 > /proc/sys/net/ipv4/ip_forward
/usr/sbin/ldirectord start 

Stop the ldirectord daemon.
#/usr/sbin/ldirectord stop 

Add the following lines to /etc/sysctl.conf.
net.ipv4.ip_forward = 1
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.eth0.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.eth0.arp_announce = 2

Execute the sysctl command to load them.
# sysctl -p 

Make ipvsadm not to automatically start at boot time
#chkconfig ipvsadm off 

Add the following lines to the /etc/hosts file.
192.168.26.201    ha1
192.168.26.202    ha2 

Configure the ha.cf file.
# cp /usr/share/doc/heartbeat-2.1.3/ha.cf /etc/ha.d/
Modify the /etc/ha.d/ha.cf file, use awk command to check the final result.
# awk ‘/^[^$]/&&/^[^#]/’ ha.cf
logfile    /var/log/ha-log
logfacility    local0
keepalive 2
deadtime 30
warntime 10
initdead 120
udpport    694
bcast    eth1        # Linux
ucast eth0 192.168.26.202
auto_failback on
node    ha1
node    ha2
ping 192.168.26.1
respawn hacluster /usr/lib64/heartbeat/ipfail
apiauth ipfail gid=haclient uid=hacluster 

Configure the haresources file and check the result.
# cp /usr/share/doc/heartbeat-2.1.3/haresources /etc/ha.d/
# awk ‘/^[^$]/&&/^[^#]/’ haresources
ha1 IPaddr::192.168.26.200/24/eth0:0/192.168.26.255 ldirectord::/etc/ha.d/conf/ldirectord.cf

Configure the authkeys file and check the result.
# cp /usr/share/doc/heartbeat-2.1.3/authkeys /etc/ha.d/
# awk ‘/^[^$]/&&/^[^#]/’ authkeys
auth 1
1 crc 

On the ha2 system
Install the heartbeat packages as my last post. 

Add the following lines to /etc/sysctl.conf.
net.ipv4.ip_forward = 1
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.eth0.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
net.ipv4.conf.eth0.arp_announce = 2 

Execute the sysctl command to load them.
# sysctl -p

Add the following lines to the /etc/hosts file.
192.168.26.201    ha1
192.168.26.202    ha2 

Copy the /etc/ha.d/ha.cf, /etc/ha.d/haresources, /etc/ha.d/authkeys and /etc/ha.d/conf/ldirectord.cf files from the ha1 system.

Modify the /etc/ha.d/ha.cf file, change ‘ucast eth0 192.168.26.202′ to ‘ucast eth0 192.168.26.201′

Start the heartbeat service
Start the heartbeat service at the ha1 and ha2 system respectively.
# service heartbeat start

Make the service to automatically start at boot time.
# chkconfig –add heartbeat 

Test
On the ha1 system
Use the ifconfig command, I could find the eth0:0.
Check the IPVS routing table, the result was as below.
# ipvsadm
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.26.200:http rr
  -> 192.168.26.211:http          Route   1      0          0        
  -> 192.168.26.212:http          Route   1      0          0

On the ha2 system
I was not able to find eth0:0 and the IPVS routing table.

On the ha1 system
Stop the heartbeat service.
# service heartbeat stop
I was not able to find eth0:0 and the IPVS routing table.

On the ha2 system
I found eth0:0 and the IPVS routing table.

On the ha1 system
Start the heartbeat service again.
I found eth0:0 and the IPVS routing table.

On the ha2 system
I was not able to find eth0:0 and the IPVS routing table.

During that period, I was able to launch the web site 192.168.26.200.