From Debuntu

Contents 1 Environment 2 Installation 3 Configuration 3.1 Enabling non-local IP binding 3.1.1 Temporarily 3.1.2 Permanently 3.2 Linux HA files 3.2.1 Authkey 3.2.2 ha.cf 3.2.2.1 LVS1 3.2.2.2 LVS2 3.2.3 haresources 4 Running the service 4.1 Starting 4.2 Runtime status 5 References

Environment

This article is going to set up a Linux High-Availability cluster using Linux HA Heartbeat between 2 Debian Lenny host.

The configuration will be made in Heartbeatv1 mode.

The host are defined as follow:

• Host 1:

Host name: lvs1

Host IP: 172.16.137.130/24 on eth0

• Host 2:

Host name: lvs2

Host IP: 172.16.137.131/24 on eth0

• Failover IP: 172.16.137.135/24

Installation

Installation is pretty straightforward and is done with APT using the following command line:

# apt-get install heartbeat

Configuration

Enabling non-local IP binding

In order to be able to bind on a IP which is not yet define on the system, we need to enable non local binding at the kernel level.

Temporarily

For testing purpose, you can use the following command:

# echo 1 > /proc/sys/net/ipv4/ip_nonlocal_bind

Permanently

To enable it permanently, you need to add to /etc/sysctl.conf :

net.ipv4.ip_nonlocal_bind = 1

and then run:

# sysctl -p

Linux HA files

There is 3 files that we will need to edit:

• /etc/ha.d/authkeys identical on both hosts

this file will contain the shared secret for nodes authentication within the cluster

• /etc/ha.d/ha.cf slightly different on host

the global cluster configuration file.

• /etc/ha.d/haresources identical on both host

Contains info on what resources are shared and who is the default master

Authkey

/etc/ha.d/authkeys will look like:

auth 1
1 md5 heartbeatmd5

Change heartbeatmd5 by any shared secret you would like.

This file will be identical on both hosts.

ha.cf

LVS1

On lvs1 /etc/ha.d/ha.cf will look like:

#       keepalive: how many seconds between heartbeats
keepalive 2
#       deadtime: seconds-to-declare-host-dead
deadtime 10
#       warntime: set late heartbeat warning time
warntime 5
#       initdead: set initial deadtime detection interval
initdead 120
#       udpport: specifies port for UDP communication
udpport 694
#       ucast: configures unicast Heartbeat communication
ucast eth0 172.16.137.131
logfacility  local0
#       auto_failback: set failback policy
auto_failback off
#       node: tells what machines are in the cluster
node lvs1 lvs2

LVS2

On lvs2 /etc/ha.d/ha.cf will look like:

#       keepalive: how many seconds between heartbeats
keepalive 2
#       deadtime: seconds-to-declare-host-dead
deadtime 10
#       warntime: set late heartbeat warning time
warntime 5
#       initdead: set initial deadtime detection interval
initdead 120
#       udpport: specifies port for UDP communication
udpport 694
#       ucast: configures unicast Heartbeat communication
ucast eth0 172.16.137.130
logfacility  local0
#       auto_failback: set failback policy
auto_failback off
#       node: tells what machines are in the cluster
node lvs1 lvs2

haresources

lvs1 172.16.137.135/24/eth0

Here we define that the IP 172.16.137.135/24 is will be running on eth0 and with preference for lvs1.

Running the service

Starting

The service can be started using:

# /etc/init.d/heartbeat start

Runtime status

You can get the status of your cluster by using the cl_status command line utility.

Here is an excerpt:

chantra@lvs1:~$ cl_status hbstatus
Heartbeat is running on this machine.
chantra@lvs1:~$ cl_status listnodes
lvs2
lvs1
chantra@lvs1:~$ cl_status nodestatus lvs1
active
chantra@lvs1:~$ cl_status nodestatus lvs2
active

References

• Linux HA Homepage
• Linux HA ha.cf wiki page
• Linux HA haresources wiki page