Building a Two-Node Linux Cluster with Heartbeat

The term "cluster" is actually not very
well defined and could mean different things to different people.
According to
Webopedia, cluster
refers to a group of disk sectors. Most Windows users are probably
familiar with lost clusters--something that can be rectified by
running the defrag utility.However, at a more advanced level in the computer industry,
cluster usually refers to a group of computers connected together
so that more computer power, e.g., more MIPS (millions instruction
per second), can be achieved or higher availability (HA) can be
obtained.Beowulf, Super Computer for the "Poor"
ApproachMost super computers in the world are built on the concept of
parallel processing--high-speed computer power is achieved by
pulling the power from each individual computer. Made by IBM, "Deep
Blue", the super computer that played chess with the world champion
Garry Kasprov, was a computer cluster that consisted of several
hundreds of RS6000s. In fact, many big time Hollywood movie
animation companies, such as Pixar, Industrial Light and Magic, use
computer clusters extensively for rendering (a process to translate
all the information such as color, movement, physical properties,
etc., into a single frame of picture).In the past, a super computer was an expensive deluxe item
that only few universities or research centers could afford.
Started at NASA, Beowulf is a project of building clusters with
"off-the-shelf" hardware (e.g., Pentium PCs) running Linux at a
very low cost.In the last several years, many universities world-wide have
set up Beowulf clusters for the purpose of scientific research or
simply for exploration of the frontier of super computer
building.High Availability (HA) ClusterClusters in this category use various technologies to gain an
extra level of reliability for a service. Companies such as Red
Hat, TurboLinux and PolyServe have cluster products that would
allow a group of computers to monitor each other; when a master
server (e.g., a web server) goes down, a secondary server will take
over the services, similar to "disk mirroring" among
servers.Simple TheoryBecause I do not have access to more than one real (or
public) IP address, I set up my two-node cluster in a private
network environment with some Linux servers and some Win9x
workstations.If you have access to three or more real/public IP addresses,
you can certainly set up the Linux cluster with real IP
addresses.In the above network diagram (fig1.gif), the Linux router is
the gateway to the Internet, and it consists of two IP addresses.
The real IP, 24.32.114.35, is attached to a network card (eth1) in
the Linux router and should be connected to either an ADSL modem or
a cable modem for internet access.The two-node Linux router consists of node1 (192.168.1.2) and
node2 (192.168.1.3). Depending on your setup, either node1 or node2
can be your primary server, and the other will be your backup
server. In this example, I will choose node1 as my primary and
node2 as my backup. Once the cluster is set, with IP aliasing (read
IP aliasing from the Linux Mini HOWTO for more detail), the primary
server will be running with an extra IP address (192.168.1.4). As
long as the primary server is up and running, services (e.g., DHCP,
DNS, HTTP, FTP, etc.) on node1 can be accessed by either
192.168.1.2 or 192.168.1.4. In fact, IP aliasing is the key concept
for setting up this two-node Linux cluster.When node1 (the primary server) goes down, node2 will be take
over all services from node1 by starting the same IP alias
(192.168.1.4) and all subsequent services. In fact, some services
can co-exist between node1 and node2 (e.g., FTP, HTTP, Samba,
etc.), however, a service such as DCHP can have only one single
running copy on the same physical segment. Likewise, we can never
have two identical IP addresses running on two different nodes in
the same network.In fact, the underlining principle of a two-node,
high-availability cluster is quite simple, and people with some
basic shell programming techniques could probably write a shell
script to build the cluster. We can set up an infinite loop within
which the backup server (node2) simply keeps pinging the primary
server, if the result is unsuccessful, and then start the floating
IP (192.168.1.4) as well as the necessary dæmons (programs
running at the background).A Two-Node Linux Cluster HOWTO with
"Heartbeat"You need two Pentium class PCs with a minimum specification
of a 100MHz CPU, 32MB RAM, one NIC (network interface card), 1G
hard drive. The two PCs need not be identical. In my experiment, I
used an AMD K6 350M Hz and a Pentium 200 MMX. I chose the AMD as my
primary server as it can complete a reboot (you need to do a few
reboots for testing) faster than the Pentium 200. With the great
support of CFSL (Computers for Schools and Libraries) in Winnipeg,
I got some 4GB SCSI hard drives as well as some Adaptec 2940 PCI
SCSI controllers. The old and almost obsolete equipment is in good
working condition and is perfect for this experiment.node1

• AMD K6 350MHz cpu
• 4G SCSI hard drive (you certainly can use IDE hard
  drive)
• 128MB RAM
• 1.44 Floppy drive
• 24x CD-ROM (not needed after installation)
• 3COM 905 NIC

node2

• Pentium 200 MMX
• 4G SCSI hard drive
• 96MB RAM
• 1.44 Floppy
• 24x CD-ROM
• 3COM 905 NIC

The Necessary SoftwareBoth node1 and node2 must have Linux installed. I chose Red
Hat and installed Red Hat 7.2 on node1 and Red Hat 6.2 on node2 (I
simply wanted to find out if we could build a cluster with
different versions of Linux installed on different nodes). Make
sure you have installed all dæmons that you want to support.
Here is my installation detail:Hard disk partitions: 128MB for swap and the rest mounted for
"/" (so that you don't need to worry about whether there is too
much or not enough for a certain subdirectory).Installed Packages:

• Apache
• FTP
• Samba
• DNS
• dhcpd (server)
• Squid

HeartbeatHeartbeat is a part of Ultra Monkey (The Linux HA Project),
and the RPM can be downloaded from
www.UltraMonkey.org.The download is small and RPM installation is smooth and
simple. However, the document or HOWTO for configuration is hard to
find and confusing. In fact, that is the reason I decided to write
this HOWTO; so that hopefully you can get your cluster setup with
less problems.Setting Up the Primary Server (node1) and the
Backup Server (node2)It is not the purpose of this article to show you how to
install Red Hat; a lot of excellent documentation can be found at
either
www.linuxdoc.org or
www.redhat.com. I will
simply include some of the most important configuration files for
your reference:

/etc/hosts
127.0.0.1       localhost
192.168.1.1     router
192.168.1.2     node1
192.168.1.3     node2

This file should be the same on both node1 and node2; you may
add any other nodes as you see fit.Check HOSTNAME (cat /etc/HOSTNAME) and
make sure it returns either node1 or node2. If not, you can use
this command (uname -n > /etc/HOSTNAME) to
fix the hostname problem.ifconfig for node1

eth0      Link encap:Ethernet  HWaddr 00:60:97:9C:52:28  
          inet addr:192.168.1.2  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:18617 errors:0 dropped:0 overruns:0 frame:0
          TX packets:14682 errors:0 dropped:0 overruns:0 carrier:0
          collisions:3 txqueuelen:100 
          Interrupt:10 Base address:0x6800 
eth0:0    Link encap:Ethernet  HWaddr 00:60:97:9C:52:28  
          inet addr:192.168.1.4  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          Interrupt:10 Base address:0x6800 
lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:3924  Metric:1
          RX packets:38 errors:0 dropped:0 overruns:0 frame:0
          TX packets:38 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0

Please notice that eth0:0 shows the IP aliasing with IP
192.168.1.4.ifconfig for node2

eth0      Link encap:Ethernet  HWaddr 00:60:08:26:B2:A4  
          inet addr:192.168.1.2  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:15673 errors:0 dropped:0 overruns:0 frame:0
          TX packets:17550 errors:0 dropped:0 overruns:0 carrier:0
          collisions:2 txqueuelen:100 
          Interrupt:10 Base address:0x6700 
lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:3924  Metric:1
          RX packets:142 errors:0 dropped:0 overruns:0 frame:0
          TX packets:142 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 

Install the Heartbeat RPMIf you are using Internet Explorer on Windows, you might have
problems accessing FTP (Netscape works much better). I suggest you
either use a command-line FTP or an FTP Windows/X Window System
client (e.g., wu_ftp) to access the FTP site of Ultra Monkey
(ftp.UltraMonkey.org).Once you log in to the FTP server of Ultra Monkey, go to pub,
then UltraMonkey and then the latest version 1.0.2 (not the beta).
The only package is heartbeat-0.4.9-1.um.1.i386.rpm; save
heartbeat-0.4.9-1.um.1.i386.rpm on your Linux box, log in as root
and install it withrpm -ivh
heartbeat-0.4.9-1.um.1.i386.rpmNull Modem Cable, Crossover Cable, Second
NICAccording to the accompanying documentation, you need to
install a second NIC on both nodes and connect them with a cross
overcable. Besides the second NIC, a null modem cable connecting
the serial (com) ports of each node is mandatory (according to the
documentation). I followed the instructions in the documentation
and installed everything. However, as I did more tests on the
cluster, I found that the null modem cable, crossover cable and the
second NIC are optional; they are nice to have but definitely
not mandatory.Configuring Heartbeat is the most important part of the whole
installation and must be set up correctly to get your cluster
working. Moreover, it should be identical on both nodes. There are
three configuration files, all stored under /etc/ha.d: ha.cf,
haresource and aythkeys.My /etc/ha.d/ha.cf

debugfile /var/log/ha-debug
#
#       File to write other messages to
#
logfile /var/log/ha-log
#
#       Facility to use for syslog()/logger 
#
logfacility     local0
#
#       keepalive: how many seconds between heartbeats
#
keepalive 2
#
#       deadtime: seconds-to-declare-host-dead
#
deadtime 10
udpport 694
#
#       What interfaces to heartbeat over?
#
udp     eth0
#
node    atm1
node    cluster1
#
# ------> end of ha.cf

Whatever is not shown above, you can simply leave as it was
(all commented out by the #). The last three options are most
important:

udp     eth0
#
node    atm1
node    cluster1

Unless you have a cross cable, you should use your eth0 (your
only NIC) for udp; the two nodes at the end of the above files must
be the same as returned by uname -n from each
node.My /etc/ha.d/haresourcesatm1 IPaddr::192.168.1.4 httpd smb
dhcpdThis is the only line you need; in the above example, I
included httpd, smb and dhcpd. You may add as many dæmons as
you want, provided they have the exact same spelling as those
dæmons under /etc/rc.d/init.dMy /etc/ha.d/authkeysYou don't need to add anything to this file, but you have to
issue the commandchmod 600 /etc/ha.d/authkeysStart the Heartbeat DaemonYou may start the dæmon withservice heartbeat startor/etc/rc.d/init.d/heartbeat startOnce heartbeat is started on both nodes, you will find that
the ifconfig from the primary server will return something
like:

node1
ifconfig for node1
eth0      Link encap:Ethernet  HWaddr 00:60:97:9C:52:28  
          inet addr:192.168.1.2  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:18617 errors:0 dropped:0 overruns:0 frame:0
          TX packets:14682 errors:0 dropped:0 overruns:0 carrier:0
          collisions:3 txqueuelen:100 
          Interrupt:10 Base address:0x6800 
eth0:0    Link encap:Ethernet  HWaddr 00:60:97:9C:52:28  
          inet addr:192.168.1.4  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          Interrupt:10 Base address:0x6800 
lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:3924  Metric:1
          RX packets:38 errors:0 dropped:0 overruns:0 frame:0
          TX packets:38 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0

When you see the line eth0:0, heartbeat is working, and you
can try to access the server by using http://192.168.1.4 and check
the log files /var/log/ha-log. Also, check the log file on node2
(192.168.1.3) and tryps -A | grep dhcpdand you should find no running dhcpd on node2.Now, the real HA test. Reboot, and then shut down the primary
server (node1: 192.168.1.2). Don't just power down the server; make
sure you issue reboot or press CTL-ALT-DEL and wait until
everything is shut down properly before you turn off your
PC.Within ten seconds, go to node2 and try ifconfig. If you can
get the IP aliasing eth0:0, you are in business and have a working
HA two-node cluster.

eth0      Link encap:Ethernet  HWaddr 00:60:08:26:B2:A4  
          inet addr:192.168.1.2  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:15673 errors:0 dropped:0 overruns:0 frame:0
          TX packets:17550 errors:0 dropped:0 overruns:0 carrier:0
          collisions:2 txqueuelen:100 
          Interrupt:10 Base address:0x6700 
eth0:0    Link encap:Ethernet  HWaddr 00:60:08:26:B2:A4  
          inet addr:192.168.1.4  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:15673 errors:0 dropped:0 overruns:0 frame:0
          TX packets:17550 errors:0 dropped:0 overruns:0 carrier:0
          collisions:2 txqueuelen:100 
          Interrupt:10 Base address:0x6700 

You can try
ps -A | grep dhcpdor you can try to release and renew the IP info on your Win9x
workstation, and you should see the new address for the dhcpd
server.Commercial ProductsCommercial products from Red Hat, TurboLinux and PolyServe
use the same concept of IP aliasing. When the primary server goes
down, the backup server will pick up the same aliasing IP so that
high availability can be achieved.The cluster product from PolyServe is very sophisticated. It
has support on SAN (server area network) and is capable of more
than two nodes. It is very easy to install and easy to configure. I
successfully configured the trial version without reading any
documentation through a windows monitoring client. However,
sophistication comes with a price tag, and the software costs more
than a thousand dollars for a two-node cluster. The 30-day trial
version cluster will stop after two hours, and it is not much fun
for testing.The cluster product from TurboLinux needs some fine-tuning.
The installation documentation is confusing (or maybe they simply
don't want people to do-it-themselves). The web configuration tool
is unstable; the cgi script will crash whenever the user clicks the
reload or refresh button. And of course, as a commercial product,
it comes with a high price tag.Linux is very stable and reliable, and it is quite common to
have our servers up and running for a few hundred days at a time.
Heartbeat works fine in my tests, and if you are looking for a
product with higher availability for a small business or education
institution, Heartbeat is definitely a perfect option.

email: leung@uwinnipeg.ca