HA Newbie

So, you're a Newbie[1] when it comes to High-Availability (HA[2])? Maybe you even feel like a clueless newbie?

Nothing wrong with that! Everybody was a newbie once, and some of us could use a clue or two on important things to this very day!

Being a newbie can be hard, and a powerful tool like Heartbeat[3] can be pretty intimidating.

So, what do you need to know to get from an intimidated newbie who doesn't know much, to someone who actually knows how to do something useful with Heartbeat[3]?

Hopefully, we'll present that here. Stay with us and see if you agree.

The method to our madness

Heartbeat[3] can be largely managed from a GUI, or it can be managed using command line tools. Because we're trying to get you to doing something useful as soon as we can, this section concentrates on the GUI way of doing things. In subsequent sections, we'll tend to present more and more of what has to be done with text files or command line tools. So, if you're a traditional UNIX command line geek, hang on, we'll get there soon enough.

We'll also try and provide examples with text, and screencast[4] video and audio as appropriate so that you stand the best chance of staying awake long enough to move on to the next stage ;-).

What do you need to do?

Here are a few things you likely need to or want to know how to do at this level, in order to get to the next level. The good news is that they're not hard to do.

1. Install Heartbeat and Start It Up ( Screencast[5])

2. Create an IP address resource ( Screencast[6], PDF[7])

3. Create a resource group with an IP address a filesystem mount, and a web server

4. Create a second resource group with a different filesystem, a database with a different IP address

5. Understand what a single point of failure is and how to avoid some of the more common ones

Once you've done these things, you'll be much less intimidated, and will have made good progress on your way to Kicking Ass[8].

Install Heartbeat

• Install from RPMs

• Compile and install from source - making an RPM along the way (optional)

• Configure a 2-node cluster - using bcast in ha.cf

  • References: ha.cf[9] documentation, authkeys[10] documentation

• Configure a 2-node cluster - using mcast, and setting autojoin any in ha.cf (optional)

  • References: ha.cf[9] documentation, authkeys[10] documentation, mcast[11] documentation, autojoin[12] documentation.

• Configuring logging
  • Learn to read the logs.
  • References: ha_logd documentation, syslog man page, syslog-ng man page.

Start It Up

• Start it up, look at log files, look at crm_mon
• References: crm_mon man page.

Create an IP address resource

IP address Screencast page[6], (screencast as PDF[7])

• Start up the GUI, create an IP address resource, start it, stop it.

  • References: LSB resource agent[13] info, OCF resource agent[14] info

Create a resource group with an IP address and a web server

• GUI already started, create a group with IP address and web server

Create a second resource group with a database with a different IP address

• GUI already started, starting from previous configuration

Single Points of Failure

• Explain what a single point of failure is
• Give a few examples
• Show how redundancy solves each of these problems

What's next?

If you can do these things, then you're not really a newbie any more, and hopefully you're a bit less intimidated as well.

There's still a lot to learn about the principles of architecting HA clusters, and also about all the amazing things Heartbeat[3] can do for you.

Now on to the next level![15]

---

This information provided courtesy of the Linux-HA project at http://linux-ha.org/