Pacemaker Explained

Configuring Pacemaker Clusters

Abstract

This document definitively explains Pacemaker’s features and capabilities, particularly the XML syntax used in Pacemaker’s Cluster Information Base (CIB).

Table of Contents

• 1. Introduction
  • 1.1. The Scope of this Document
  • 1.2. What Is Pacemaker?
    • 1.2.1. Cluster Architecture
    • 1.2.2. Pacemaker Architecture
    • 1.2.3. Node Redundancy Designs
• 2. Host-Local Configuration
  • 2.1. Configuration Value Types
    • 2.1.1. Scores
  • 2.2. Local Options
• 3. Cluster-Wide Configuration
  • 3.1. Configuration Layout
  • 3.2. Option Precedence
  • 3.3. CIB Properties
  • 3.4. Cluster Options
• 4. Nodes
  • 4.1. Cluster nodes
    • 4.1.1. Host Clock Considerations
  • 4.2. Pacemaker Remote nodes
  • 4.3. Defining a Node
    • 4.3.1. Where Pacemaker Gets the Node Name
  • 4.4. Quorum-only Nodes
  • 4.5. Node Attributes
    • 4.5.1. Setting and querying node attributes
    • 4.5.2. Special node attributes
  • 4.6. Tracking Node Health
    • 4.6.1. Node Health Attributes
    • 4.6.2. Node Health Strategy
    • 4.6.3. Exempting a Resource from Health Restrictions
    • 4.6.4. Configuring Node Health Agents
• 5. Resources
  • 5.1. Resource Standards
    • 5.1.1. Open Cluster Framework
    • 5.1.2. Systemd
    • 5.1.3. Linux Standard Base
    • 5.1.4. System Services
    • 5.1.5. STONITH
  • 5.2. Resource Properties
  • 5.3. Resource Options
    • 5.3.1. Resource Meta-Attributes
    • 5.3.2. Setting Global Defaults for Resource Meta-Attributes
    • 5.3.3. Resource Instance Attributes
  • 5.4. Pacemaker Remote Resources
    • 5.4.1. Remote nodes
    • 5.4.2. Guest Nodes
    • 5.4.3. Removing Pacemaker Remote Nodes
• 6. Resource Operations
  • 6.1. Operation Properties
  • 6.2. Monitoring Resources for Failure
  • 6.3. Custom Recurring Operations
  • 6.4. Setting Global Defaults for Operations
  • 6.5. When Implicit Operations Take a Long Time
  • 6.6. Multiple Monitor Operations
  • 6.7. Disabling a Monitor Operation
  • 6.8. Handling Resource Failure
    • 6.8.1. Failure Counts
    • 6.8.2. Failure Response
  • 6.9. Reloading an Agent After a Definition Change
  • 6.10. Migrating Resources
• 7. Resource Constraints
  • 7.1. Deciding Which Nodes a Resource Can Run On
    • 7.1.1. Location Properties
    • 7.1.2. Asymmetrical “Opt-In” Clusters
    • 7.1.3. Symmetrical “Opt-Out” Clusters
    • 7.1.4. What if Two Nodes Have the Same Score
    • 7.1.5. Specifying locations using pattern matching
  • 7.2. Specifying the Order in which Resources Should Start/Stop
    • 7.2.1. Ordering Properties
    • 7.2.2. Optional and mandatory ordering
    • 7.2.3. Symmetric and asymmetric ordering
  • 7.3. Placing Resources Relative to other Resources
    • 7.3.1. Colocation Properties
    • 7.3.2. Mandatory Placement
    • 7.3.3. Advisory Placement
    • 7.3.4. Colocation by Node Attribute
    • 7.3.5. Colocation Influence
  • 7.4. Resource Sets
    • 7.4.1. Anti-colocation Chains
  • 7.5. Ordering Sets of Resources
    • 7.5.1. Ordered Set
    • 7.5.2. Ordering Multiple Sets
    • 7.5.3. Resource Set OR Logic
  • 7.6. Colocating Sets of Resources
  • 7.7. External Resource Dependencies
• 8. Fencing
  • 8.1. What Is Fencing?
  • 8.2. Why Is Fencing Necessary?
  • 8.3. Fence Devices
  • 8.4. Fence Agents
  • 8.5. When a Fence Device Can Be Used
  • 8.6. Limitations of Fencing Resources
  • 8.7. Special Meta-Attributes for Fencing Resources
  • 8.8. Special Instance Attributes for Fencing Resources
  • 8.9. Default Check Type
  • 8.10. Unfencing
  • 8.11. Fencing and Quorum
  • 8.12. Fencing Timeouts
  • 8.13. Fence Devices Dependent on Other Resources
  • 8.14. Configuring Fencing
    • 8.14.1. Example Fencing Configuration
  • 8.15. Fencing Topologies
    • 8.15.1. Example Dual-Layer, Dual-Device Fencing Topologies
  • 8.16. Remapping Reboots
• 9. Collective Resources
  • 9.1. Groups - A Syntactic Shortcut
    • 9.1.1. Group Properties
    • 9.1.2. Group Options
    • 9.1.3. Group Instance Attributes
    • 9.1.4. Group Contents
    • 9.1.5. Group Constraints
    • 9.1.6. Group Stickiness
  • 9.2. Clones - Resources That Can Have Multiple Active Instances
    • 9.2.1. Anonymous versus Unique Clones
    • 9.2.2. Promotable clones
    • 9.2.3. Clone Properties
    • 9.2.4. Clone Options
    • 9.2.5. Clone Contents
    • 9.2.6. Clone Instance Attribute
    • 9.2.7. Clone Constraints
    • 9.2.8. Clone Stickiness
    • 9.2.9. Monitoring Promotable Clone Resources
    • 9.2.10. Determining Which Instance is Promoted
  • 9.3. Bundles - Containerized Resources
    • 9.3.1. Bundle Prerequisites
    • 9.3.2. Bundle Properties
    • 9.3.3. Bundle Container Properties
    • 9.3.4. Bundle Network Properties
    • 9.3.5. Bundle Storage Properties
    • 9.3.6. Bundle Primitive
    • 9.3.7. Bundle Node Attributes
    • 9.3.8. Bundle Meta-Attributes
    • 9.3.9. Limitations of Bundles
• 10. Utilization and Placement Strategy
  • 10.1. Utilization attributes
  • 10.2. Placement Strategy
  • 10.3. How Multiple Capacities Combine
  • 10.4. Order of Resource Assignment
  • 10.5. Limitations
• 11. Rules
  • 11.1. Rule Options
  • 11.2. Rule Conditions and Contexts
  • 11.3. Date/Time Expressions
    • 11.3.1. Date Specifications
    • 11.3.2. Durations
    • 11.3.3. Example Date/Time Expressions
  • 11.4. Node Attribute Expressions
  • 11.5. Resource Type Expressions
    • 11.5.1. Example Resource Type Expressions
  • 11.6. Operation Type Expressions
    • 11.6.1. Example Operation Type Expressions
  • 11.7. Using Rules to Determine Resource Location
    • 11.7.1. Location Rules Based on Other Node Properties
    • 11.7.2. Using score-attribute Instead of score
    • 11.7.3. Specifying location scores using pattern submatches
  • 11.8. Using Rules to Define Options
    • 11.8.1. Using Rules to Control Resource Options
    • 11.8.2. Using Rules to Control Resource Defaults
    • 11.8.3. Using Rules to Control Cluster Options
• 12. Access Control Lists (ACLs)
  • 12.1. ACL Prerequisites
  • 12.2. ACL Configuration
  • 12.3. ACL Roles
  • 12.4. ACL Targets and Groups
  • 12.5. ACLs and Pacemaker Remote Nodes
  • 12.6. ACL Examples
  • 12.7. ACL Limitations
    • 12.7.1. Actions performed via IPC rather than the CIB
    • 12.7.2. ACLs and Pacemaker Remote
• 13. Alerts
  • 13.1. Alert Agents
  • 13.2. Alert Recipients
  • 13.3. Alert Meta-Attributes
  • 13.4. Alert Instance Attributes
  • 13.5. Alert Filters
• 14. Reusing Parts of the Configuration
  • 14.1. Reusing Resource Definitions
    • 14.1.1. Configuring Resources with Templates
    • 14.1.2. Using Templates in Constraints
    • 14.1.3. Using Templates in Resource Sets
  • 14.2. Reusing Rules, Options and Sets of Operations
  • 14.3. Tagging Configuration Elements
    • 14.3.1. Configuring Tags
    • 14.3.2. Using Tags in Constraints and Resource Sets
    • 14.3.3. Filtering With Tags
• 15. Status
  • 15.1. Node State
  • 15.2. Transient Node Attributes
  • 15.3. Node History
    • 15.3.1. Resource History
    • 15.3.2. Action History
    • 15.3.3. Simple Operation History Example
    • 15.3.4. Complex Operation History Example
• 16. Multi-Site Clusters and Tickets
  • 16.1. Challenges for Multi-Site Clusters
  • 16.2. Conceptual Overview
    • 16.2.1. Ticket
    • 16.2.2. Dead Man Dependency
    • 16.2.3. Cluster Ticket Registry
    • 16.2.4. Configuration Replication
  • 16.3. Configuring Ticket Dependencies
  • 16.4. Managing Multi-Site Clusters
    • 16.4.1. Granting and Revoking Tickets Manually
    • 16.4.2. Granting and Revoking Tickets via a Cluster Ticket Registry
    • 16.4.3. General Management of Tickets
  • 16.5. For more information
• 17. Sample Configurations
  • 17.1. Empty
  • 17.2. Simple
  • 17.3. Advanced Configuration

Index

• Index

• Search Page