The Novice

Since you're not a complete newbie[1], you can actually do something useful with Heartbeat[2]. Now that you know enough to be dangerous, let's expand on that into more real-world tasks that you will likely want to do at the next level of difficulty, without becoming a danger to yourself or others.

And, now we also begin to add some information on some of the general principles of how to architect an HA cluster, show you how the XML configuration behind the scenes works, and some of the command line tools you need to know about.

We may need to split this into a few levels... There are just so many things to think about... Here are a few - in no particular order...

Heartbeat-specific topics

• Heartbeat "principles of operation" (POO)
• Heartbeat architectural overview
• What is the CIB?
• Global CIB parameters (2 parts?)
• Ordering Dependencies
• Colocation Dependencies
• Use of resource-stickiness
• Use of resource-failure-stickiness

• XML <rule>s

• crm_verify
• dumping the CIB using cibadmin
• editing the CIB using cibadmin
• Node Attributes
• Ping nodes
• pingd

• STONITH[3]

• Global Configuration settings
• UUIDs

• restoring Heartbeat[2] onto a machine where the configuration was wiped out

• Ucast
• bcast port numbers
• possible mcast problems
• serial ???
• Resource agent parameters and attributes - what's the difference?
• Common Resource Agents overview: IPaddr[2], Filesystem, Apache, etc.
• master/slave resources
• configuring and using DRBD
• clone resources
• clone STONITH resources
• Writing OCF (Open Cluster Framework) resource agents
  • OCF Metadata
  • OCF return codes
  • Testing resource agents
  • Mention OCF_RESKEY_foobar variables
• Converting an R1 (haresources) configuration over to an R2 (CRM/CIB) configuration

• CIB Idioms[4]

  • master / slave
  • others :-D
• Common problems
  • Firewalls: yes you probably have one, I don't care what you think
  • Testing the wrong way -- and the right way

General HA topics

• ntp - synchronizing clocks

• split-brain[5], Fencing[6] and Quorum - basic overview[7]

• Redundancy, Redundancy, and Redundancy
• Data/state sharing methods: replication
• Data/state sharing methods: shared storage
• Network Redundancy
• How to tell when you're in a DR situation

---

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