vagrant up lb1 lb2 webserver1 webserver2 webserver3
watch -n 1 curl -I 192.168.10.80
watch -n 1 curl -I 192.168.10.80
- stop
lb1
:vagrant halt lb1
- start
lb1
:vagrant up lb1
- stop
lb2
:vagrant halt lb1
- start
lb2
:vagrant up lb1
Both IPs are available during this test. The failover need 1-2 seconds.
- 2 Loadbalancer
- lb1:
192.168.10.50
- lb2:
192.168.10.51
- 3 Weberserver
- webserver1:
192.168.10.21
- webserver2:
192.168.10.22
- webserver3:
192.168.10.23
192.168.10.80
and 192.168.10.90
are floating IPs between lb1
and lb2
lb1
is the master for 192.168.10.90
and lb2
for 192.168.10.80
.
Use conntrackd for share the TCP states between both LBs -> no downtime during the failover