Load balancing is a way to distribute workloads across multiple computing resources such that large, resource-intensive tasks may be safely and reliable completed with more efficiency and speed than if just one machine were to perform the tasks.<\/p>\n
Keepalived is a free and open source load balancing solution for Linux systems. This guide will go over the installation process for Keepalived on Ubuntu 14.04 and how you can use it for load balancing on your Linux clusters.<\/p>\n
Before you begin to follow the steps in this guide, make sure that you meet these requirements: For the purposes of this guide, we’ll be working with a public network of 173.209.49.66\/29 and a class A private network, or LAN, of 10.119.0.0\/24.<\/p>\n For your reference, here are the servers, or load balancers, we’ll be working with, along with their respective public and private IP addresses. Where necessary, remember to replace with the IP addresses for your own servers.<\/p>\n LB1 LB2 The load balancers will make use of a “floating IP”, and we’ll configure active and passive redundancy as well.<\/p>\n Floating The first task is to ensure that the systems of both servers are fully up to date.<\/p>\n If it’s installed, disable Ubuntu’s default firewall.<\/p>\n The next step is to install Keepalived and all necessary dependencies.<\/p>\n Use this command to activate Keepalived on boot. We’ll also enable the ipvsadm kernel module.<\/p>\n Now we’ll have to configure Keepalived for our setup.<\/p>\n Here’s the alterations for the LB1 server.<\/p>\n And these alterations will be applied to the LB2 server.<\/p>\n The “virtual_router_id” setting must be unique for each of the defined VRRP instances, and it should also be unique within your VLAN. Make sure you’re not using the same ID on any two clusters connected via the same physical switch or VLAN. Naturally, this ID needs to match on both LB1 and LB2 for the same VRRP instance. Valid values are from 0 to 255.<\/p>\n Now, enable nf_conntrack and we can proceed to sysctl configuration.<\/p>\n Alter the configuration so it matches the below:<\/p>\n Now, apply the changes. Finally, we can start up Keepalived.<\/p>\n Let’s verify that Keepalived is running in the expected manner. First we should verify that both of the floating IPs are assigned to the first keepalived instance.<\/p>\n Using the command ip addr show, you can see if the IPs are present:<\/p>\n If everything’s been set up correctly, you’ll see 173.209.49.70 and 10.119.0.10 on LB1. If you shut down keepalived on LB1, those same IP addresses will appear on the second server.<\/p>\n After shutting down keepalived on LB1, go on the second server and see if it indeed has those IP addresses assigned:<\/p>\n Finally, make sure that the backends are well specified within keepalived’s config:<\/p>\n Now that you’ve completed this guide, you should be able to set up Keepalived on your own clusters. Keepalived is scalable, so you can try it out on clusters of any size. If you found this article helpful, feel free to share it with your friends and let us know in the comments below! <\/p>\n","protected":false},"excerpt":{"rendered":" How to Achieve High Availability Load Balancing with Keepalived on Ubuntu 14 Load balancing is a way to distribute workloads across multiple computing resources such that large, resource-intensive tasks may be safely and reliable completed with more efficiency and speed than if just one machine were to perform the tasks. Keepalived is a free and<\/p>\n","protected":false},"author":3,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[75],"tags":[],"class_list":["post-3140","post","type-post","status-publish","format-standard","hentry","category-ha-clustering","operating_system-ubuntu-14-04"],"yoast_head":"\n
\n\u2022 Two servers (Cloud Server<\/b><\/a> or Dedicated Server<\/b><\/a>), each running a fresh installation of Ubuntu 14.04. We will call these servers LB1 and LB2 below
\n\u2022 Both servers connected to the same LAN
\n\u2022 Root access to both servers<\/p>\nTutorial<\/h2>\n
\nPublic:173.209.49.66
\nPrivate:10.119.0.1<\/p>\n
\nPublic:173.209.49.67
\nPrivate:10.119.0.2<\/p>\n
\nPublic:173.209.49.70
\nPrivate:10.119.0.10<\/p>\napt-get update
\napt-get -y upgrade<\/code><\/p>\nufw disable<\/code><\/p>\napt-get install linux-headers-$(uname -r) keepalived<\/code><\/p>\nupdate-rc.d keepalived defaults
\nmodprobe ip_vs<\/code><\/p>\necho \"\" > \/etc\/keepalived\/keepalived.conf
\nnano \/etc\/keepalived\/keepalived.conf<\/code><\/p>\nvrrp_instance VI_LOCAL {
\n interface eth1
\n state MASTER
\n virtual_router_id 51
\n priority 101
\n virtual_ipaddress {
\n 10.119.0.10
\n }
\n track_interface {
\n eth0
\n eth1
\n }
\n}
\nvrrp_instance VI_PUB {
\n interface eth0
\n state MASTER
\n virtual_router_id 52
\n priority 101
\n virtual_ipaddress {
\n 173.209.49.70
\n }
\n track_interface {
\n eth0
\n eth1
\n }
\n}
\nvirtual_server 173.209.49.70 443 {
\n delay_loop 4
\n lb_algo sh # source hash
\n lb_kind NAT
\n protocol TCP
\n real_server 10.119.0.100 443 {
\n weight 1
\n TCP_CHECK {
\n connect_timeout 15
\n nb_get_retry 3
\n delay_before_retry 2
\n }
\n }
\n real_server 10.119.0.101 443 {
\n weight 1
\n TCP_CHECK {
\n connect_timeout 15
\n nb_get_retry 3
\n delay_before_retry 2
\n }
\n }
\n}
\nvirtual_server 173.209.49.70 80 {
\n delay_loop 4
\n lb_algo wrr # weighted round robin
\n lb_kind NAT
\n protocol TCP
\n real_server 10.119.0.100 80 {
\n weight 1
\n TCP_CHECK {
\n connect_timeout 15
\n nb_get_retry 3
\n delay_before_retry 2
\n }
\n }
\n real_server 10.119.0.101 80 {
\n weight 1
\n TCP_CHECK {
\n connect_timeout 15
\n nb_get_retry 3
\n delay_before_retry 2
\n }
\n }
\n}<\/code><\/p>\nvrrp_instance VI_LOCAL {
\n interface eth1
\n state BACKUP
\n virtual_router_id 51
\n priority 100
\n virtual_ipaddress {
\n 10.119.0.10
\n }
\n track_interface {
\n eth0
\n eth1
\n }
\n}
\nvrrp_instance VI_PUB {
\n interface eth0
\n state BACKUP
\n virtual_router_id 52
\n priority 100
\n virtual_ipaddress {
\n 173.209.49.70
\n }
\n track_interface {
\n eth0
\n eth1
\n }
\n}
\nvirtual_server 173.209.49.70 443 {
\n delay_loop 4
\n lb_algo sh # source hash
\n lb_kind NAT
\n protocol TCP
\n real_server 10.119.0.100 443 {
\n weight 1
\n TCP_CHECK {
\n connect_timeout 15
\n nb_get_retry 3
\n delay_before_retry 2
\n }
\n }
\n real_server 10.119.0.101 443 {
\n weight 1
\n TCP_CHECK {
\n connect_timeout 15
\n nb_get_retry 3
\n delay_before_retry 2
\n }
\n }
\n}
\nvirtual_server 173.209.49.70 80 {
\n delay_loop 4
\n lb_algo wrr # weighted round robin
\n lb_kind NAT
\n protocol TCP
\n real_server 10.119.0.100 80 {
\n weight 1
\n TCP_CHECK {
\n connect_timeout 15
\n nb_get_retry 3
\n delay_before_retry 2
\n }
\n }
\n real_server 10.119.0.101 80 {
\n weight 1
\n TCP_CHECK {
\n connect_timeout 15
\n nb_get_retry 3
\n delay_before_retry 2
\n }
\n }
\n}<\/code><\/p>\nmodprobe nf_conntrack
\nnano \/etc\/sysctl.conf<\/code><\/p>\nnet.ipv4.ip_forward = 1
\nnet.ipv4.ip_nonlocal_bind = 1
\nnet.nf_conntrack_max = 1000000<\/code><\/p>\n
\nsysctl -p<\/code><\/p>\nservice keepalived start<\/code><\/p>\nroot@lb1:\/etc# ip addr show<\/code><\/p>\n1: lo:
\n link\/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
\n inet 127.0.0.1\/8 scope host lo
\n valid_lft forever preferred_lft forever
\n inet6 ::1\/128 scope host
\n valid_lft forever preferred_lft forever
\n2: eth0:
\n link\/ether 00:50:56:8e:e4:2f brd ff:ff:ff:ff:ff:ff
\n inet 173.209.49.66\/29 brd 173.209.49.71 scope global eth0
\n valid_lft forever preferred_lft forever
\n inet 173.209.49.70\/32 scope global eth0
\n valid_lft forever preferred_lft forever
\n inet6 fe80::250:56ff:fe8e:e42f\/64 scope link
\n valid_lft forever preferred_lft forever
\n3: eth1:
\n link\/ether 00:50:56:8e:ea:2d brd ff:ff:ff:ff:ff:ff
\n inet 10.119.0.1\/24 brd 10.119.0.255 scope global eth1
\n valid_lft forever preferred_lft forever
\n inet 10.119.0.10\/32 scope global eth1
\n valid_lft forever preferred_lft forever
\n inet6 fe80::250:56ff:fe8e:ea2d\/64 scope link
\n valid_lft forever preferred_lft forever<\/code><\/p>\nroot@lb1:\/etc# systemctl stop keepalived<\/code><\/p>\nroot@lb2:~# ip addr show<\/code><\/p>\n1: lo:
\n link\/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
\n inet 127.0.0.1\/8 scope host lo
\n valid_lft forever preferred_lft forever
\n inet6 ::1\/128 scope host
\n valid_lft forever preferred_lft forever
\n2: eth0:
\n link\/ether 00:50:56:8e:ae:b8 brd ff:ff:ff:ff:ff:ff
\n inet 173.209.49.67\/29 brd 173.209.49.71 scope global eth0
\n valid_lft forever preferred_lft forever
\n inet 173.209.49.70\/32 scope global eth0
\n valid_lft forever preferred_lft forever
\n inet6 fe80::250:56ff:fe8e:aeb8\/64 scope link
\n valid_lft forever preferred_lft forever
\n3: eth1:
\n link\/ether 00:50:56:8e:ed:ba brd ff:ff:ff:ff:ff:ff
\n inet 10.119.0.2\/24 brd 10.119.0.255 scope global eth1
\n valid_lft forever preferred_lft forever
\n inet 10.119.0.10\/32 scope global eth1
\n valid_lft forever preferred_lft forever
\n inet6 fe80::250:56ff:fe8e:edba\/64 scope link
\n valid_lft forever preferred_lft forever<\/code><\/p>\nroot@lb1:\/etc# ipvsadm<\/code><\/p>\nIP Virtual Server version 1.2.1 (size=4096)
\nProt LocalAddress:Port Scheduler Flags
\n -> RemoteAddress:Port Forward Weight ActiveConn InActConn
\nTCP 173.209.49.70:http wrr
\n -> 10.119.0.100:http Masq 1 0 0
\n -> 10.119.0.101:http Masq 1 0 0
\nTCP 173.209.49.70:https sh
\n -> 10.119.0.100:https Masq 1 0 0
\n -> 10.119.0.101:https Masq 1 0 0<\/code><\/p>\nConclusion<\/h2>\n