Setting up a Load-Balancing Cluster with Heartbeat and Ldirectord

We will be setting up a load balancer with Heartbeat and Ldirector on CentOS 6.

Note

Ldirectord has been removed from RHEL 6 default repository and replaced with Piranha.

As of RHEL 6.6, Red Hat provides support for HAProxy and keepalived in addition to the Piranha load balancing software.

Piranha has been removed from RHEL 7 default repository and replaced with HAProxy and keepalived.

Software

Software used in this article:

  1. CentOS 6.7
  2. Heartbeat 3.0.4
  3. Ldirectord 3.9.5
  4. Ipvsadm 1.26

Networking and IP Addresses

Our network is set up as follows:

  1. 10.8.8.0/24 – LAN with access to the Internet.

Hostnames and roles of the virtual machines we are going to use:

  1. lvs-ldir01 – the active ldirectord router with heartbeat,
  2. lvs-ldir02 – the backup ldirectord router with heartbeat,
  3. lvs-ldir03/lvs-ldir04 – real servers, both running a pre-configured Apache webserver.

See the schema below for more information.

Direct Routing Real Server Node Configuration

On each real server node (lvs-ldir03 and lvs-ldir04), run the following command for the VIP 10.8.8.45 and protocol combination intended to be serviced for the real server:

# iptables -t nat -A PREROUTING -p tcp -d 10.8.8.45 --dport 80 -j REDIRECT

The command above will cause the real servers to process packets destined for the VIP and port that they are given. Make sure that firewall changes are saved and restored after a restart:

# service iptables save

Ldirectord Setup

Linux Director Daemon (ldirectord) is a background process computer program used to monitor and administer real servers in the Linux Virtual Server (LVS) cluster. Ldirectord monitors the health of the real servers by periodically requesting a known URL and checking that the response contains an expected string.

While ldirectord is used to monitor and administer real servers in the LVS cluster, heartbeat is used as the fail-over monitor for the load balancers (ldirectord).

Installation

The ldirectord installation steps must be repeated on both routers, the lvs-ldir01 and the lvs-ldir02.

Install dependencies for ldirectord:

# yum install -y ipvsadm perl-libwww-perl perl-IO-Socket-INET6 perl-Net-SSLeay perl-MailTools

Ensure the IP Virtual Server Netfilter kernel module is loaded:

# modprobe ip_vs

A fairly old (2010-03-21) heartbeat-ldirectord v2.1.4 package is available from epel5 repositories (direct link).

We are going to use an ldirectord v3.9.5 package which was released on 2013-07-26.

Download it from here and install.

# rpm -Uvh ldirectord-3.9.5-3.1.x86_64.rpm

Disable on boot as it will be managed by heartbeat:

# chkconfig ldirectord off

Configure

Create the /etc/ha.d/ldirectord.cf file and add the following content:

# Global Directives
checkinterval = 5
checktimeout = 10
autoreload = no
logfile = "/var/log/ldirectord.log"
#logfile="local0"
quiescent = yes

#
# virtual = x.y.z.w:p
# protocol = tcp|udp
# scheduler = rr|wrr|lc|wlc
# real = x.y.z.w:p gate|masq|ipip [weight]
#

# Virtual Server for HTTP
virtual = 10.8.8.45:80
	real = 10.8.8.48:80 gate 1
	real = 10.8.8.49:80 gate 1
	service = http
	protocol = tcp
	request = "lbcheck.html"
	receive = "found"
	scheduler = wlc
	checktype = negotiate

The directives under “virtual=” have to start with a [TAB], not white space. The lbcheck.html file with the content of “found” must be present on the real webservers.

See “man ldirectord” for configuration directives not covered here.

Test ldirectord on the lvs-ldir01 router:

# /etc/init.d/ldirectord start

We should see something like this:

# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port   Forward Weight ActiveConn InActConn
TCP  10.8.8.45:80 wlc
  -> 10.8.8.48:80         Route   1      0          0         
  -> 10.8.8.49:80         Route   1      0          0

Let us stop the ldirectord service:

 # /etc/init.d/ldirectord stop

Heartbeat Setup

Heartbeat is a basic high-availability subsystem for Linux-HA.

Instalation

The heartbeat installation steps must be repeated on both routers, the lvs-ldir01 and the lvs-ldir02.

You can, for example, configure one router node and then rsync config to the other one:

# rsync -ah --progress /etc/ha.d/ [email protected]:/etc/ha.d/

Install heartbeat:

# yum install -y heartbeat

Enable on boot:

# chkconfig heartbeat on

Enable forwarding:

# sed -i 's/net.ipv4.ip_forward = 1/net.ipv4.ip_forward = 0' /etc/sysctl.conf
# sysctl -p

Configuration

Create the /etc/ha.d/ha.cf file and add the following content:

logfile /var/log/ha.log
logfacility local0
debug 0

# Disable the Pacemaker cluster manager
# crm off|on|respawn
crm off

# Interval between heartbeat packets
keepalive 2
# How quickly Heartbeat should decide that a node in a cluster is dead
deadtime 6

# Which port Heartbeat should listen on
udpport 694
# Which interfaces Heartbeat sends UDP broadcast traffic on
bcast eth0

# Automatically fail back to a "primary" node (deprecated)
auto_failback off

# What machines are in the cluster. Use "uname -n"
node lvs-ldir01.hl.local
node lvs-ldir02.hl.local

Configure iptables firewall to allow heartbeat and HTTP traffic:

# iptables -A INPUT -s 10.8.8.0/24 -p udp --dports 694 -j ACCEPT
# iptables -A INPUT -s 10.0.0.0/8 -p tcp --dports 80 -j ACCEPT

Save firewall rules:

# service iptables save

The ldirectored can be easily started and stopped by heartbeat. Put the ldirectord under the /etc/ha.d/resource.d/ directory, then we can add a line in the /etc/ha.d/haresources like:

lvs-ldir01.hl.local IPaddr::10.8.8.45 ldirectord::ldirectord.cf

Now, start the heartbeat service on both routers, the lvs-ldir01and the lvs-ldir02:

# /etc/init.d/heartbeat start

At this point, the virtual IP should be assigned to the eth0 interface on the primary router:

# ip ad show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 08:00:27:ff:46:00 brd ff:ff:ff:ff:ff:ff
    inet 10.8.8.46/24 brd 10.8.8.255 scope global eth0
    inet 10.8.8.45/24 brd 10.8.8.255 scope global secondary eth0
    inet6 fe80::a00:27ff:feff:4600/64 scope link 
       valid_lft forever preferred_lft forever
# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port   Forward Weight ActiveConn InActConn
TCP  10.8.8.45:80 wlc
  -> 10.8.8.48:80         Route   1      0          15         
  -> 10.8.8.49:80         Route   1      0          16

Failover Test

Stop the heartbear service on the primary router node lvs-ldir01:

[[email protected] ~]# /etc/init.d/heartbeat stop

The lvs-ldir02 node should become active:

[[email protected] ~]# ip ad show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 08:00:27:ff:47:00 brd ff:ff:ff:ff:ff:ff
    inet 10.8.8.47/24 brd 10.8.8.255 scope global eth0
    inet 10.8.8.45/24 brd 10.8.8.255 scope global secondary eth0
    inet6 fe80::a00:27ff:feff:4700/64 scope link 
       valid_lft forever preferred_lft forever

Excerpts from the /var/log/ha.log:

lvs-ldir02.hl.local heartbeat: [4808]: info: mach_down takeover complete.
lvs-ldir02.hl.local heartbeat: [4808]: WARN: node lvs-ldir01.hl.local: is dead
lvs-ldir02.hl.local heartbeat: [4808]: info: Dead node lvs-ldir01.hl.local gave up resources.
lvs-ldir02.hl.local heartbeat: [4808]: info: Link lvs-ldir01.hl.local:eth0 dead.

Now, if we start the heartbeat service on the lvs-ldir01 node, resources will be transferred from the lvs-ldir02 back to the lvs-ldir01 due to “auto_failback on”.

Troubleshooting

Check logs /var/log/ha.log and /var/log/ldirectord.log.

References

https://www.suse.com/communities/blog/load-balancing-howto-lvs-ldirectord-heartbeat-2/
http://www.linux-ha.org/doc/man-pages/re-hacf.html
http://www.linuxvirtualserver.org/docs/ha/heartbeat_ldirectord.html

2 thoughts on “Setting up a Load-Balancing Cluster with Heartbeat and Ldirectord

Leave a Reply

Your email address will not be published. Required fields are marked *