Data ONTAP 7.3 Active/Active Configuration Guide
Copyright information
Trademark information
About this guide
Audience
Terminology
Command, keyboard, and typographic conventions
Special messages
Active/active configurations and requirements
Overview of active/active configurations
What an active/active configuration is
Benefits of active/active configuration
Characteristics of nodes in an active/active configuration
Using best practices when deploying active/active configurations
Comparison of active/active configuration types
Standard active/active configurations
How Data ONTAP works with standard active/active configurations
Standard active/active configuration diagram
Setup requirements and restrictions for standard active/active configurations
Configuration variations for standard active/active configurations
Understanding mirrored active/active configurations
Advantages of mirrored active/active configurations
Setup requirements and restrictions for mirrored active/active configurations
Configuration variations for mirrored active/active configurations
Understanding stretch MetroClusters
Continued data service after loss of one node with MetroCluster
Advantages of stretch MetroCluster configurations
Stretch MetroCluster configuration
Stretch MetroCluster configuration in dual-controller systems
How Data ONTAP works with stretch MetroCluster configurations
Stretch MetroCluster and disk ownership
Setup requirements and restrictions for stretch MetroCluster configurations
Configuration variations for stretch MetroCluster configurations
Understanding fabric-attached MetroClusters
Fabric-attached MetroClusters use Brocade Fibre Channel switches
Advantages of fabric-attached MetroCluster configurations
Fabric-attached MetroCluster configuration
Fabric-attached MetroCluster configuration with dual-controller systems
How Data ONTAP works with fabric-attached MetroCluster configurations
Setup requirements and restrictions for fabric-attached MetroClusters
Configuration limitations for fabric-attached MetroCluster configurations
Configuration variations for fabric-attached MetroCluster configurations
Active/active installation
System cabinet or equipment rack installation
Active/active configurations in an equipment rack
Active/active configurations in a system cabinet
Required documentation, tools, and equipment
Required documentation
Required tools
Required equipment
Preparing your equipment
Installing the nodes in equipment racks
Installing the nodes in a system cabinet
Cabling procedures for standard or mirrored active/active configurations
Systems with two controllers in the same chassis
Which Fibre Channel ports to use for an active/active configuration
Cabling a standard active/active configuration
Cabling Node A for a standard active/active configuration
Cabling Node B for a standard active/active configuration
Cabling the cluster interconnect for a standard active/active configuration
Cabling a mirrored active/active configuration
Mirrored active/active configurations and disk pools
Which Fibre Channel ports to use for an active/active configuration
Creating your port list for mirrored active/active configurations
Using the mirrored active/active configuration diagrams
Cabling the channel A loops for a mirrored active/active configuration
Cabling the channel B loops for a mirrored active/active configuration
Cabling the cluster interconnect for a mirrored active/active configuration
About using a UPS with standard or mirrored active/active configurations
MetroCluster installation
Required documentation, tools, and equipment
Required documentation
Required tools
Required equipment
MetroCluster and software-based disk ownership
Converting an active/active configuration to a fabric-attached MetroCluster
Upgrading an existing fabric-attached MetroCluster
Cabling a stretch MetroCluster
Cabling a stretch MetroCluster between dual-controller systems
Cabling a fabric-attached MetroCluster
Planning the fabric-attached MetroCluster installation
Configuration differences for fabric-attached MetroClusters on dual-controller systems
Configuring the switches
Switch bank rules and virtual channel rules
Cabling Node A
Cabling the controller to the switches
If you have hardware-based disk ownership
If you have software-based disk ownership
Tips for controller-to-switch connections
Cabling the disk shelves to the switches
If you have hardware-based disk ownership
If you have software-based disk ownership
Tips for disk shelf-to-switch connections
Cabling the FC-VI adapter and inter-switch link
If you have hardware-based disk ownership
If you have software-based disk ownership
Tips for cluster interconnect connections
Cabling Node B
Cabling the Node B controller to the switches
If you have hardware-based disk ownership
If you have software-based disk ownership
Tips for controller-to-switch connections
Cabling the disk shelves to the switches
If you have hardware-based disk ownership
If you have software-based disk ownership
Tips for disk shelf-to-switch connections
Cabling the FC-VI adapter and inter-switch link
If you have hardware-based disk ownership
If you have software-based disk ownership
Tips for cluster interconnect connections
Assigning disk pools (if you have software-based disk ownership)
Verifying disk paths
If you have hardware-based disk ownership
If you have software-based disk ownership
About using a UPS with MetroCluster configurations
Configuring an active/active configuration
Bringing up the active/active configuration
Considerations for active/active configuration setup
Configuring shared interfaces with setup
Configuring dedicated interfaces with setup
Configuring standby interfaces with setup
Enabling licenses
Setting options and parameters
Option types for active/active configurations
Setting matching node options
Parameters that must be the same on each node
Disabling the change_fsid option in MetroCluster configurations
Configuration of the hw_assist option
Hardware-assisted takeover
Disabling and enabling the hardware-assisted takeover option
Setting the partner address for hardware-assisted takeover
Setting the partner port for hardware-assisted takeover
Configuration of network interfaces
What the networking interfaces do
Configuring network interfaces for the active/active configuration
Testing takeover and giveback
Management of an active/active configuration
How takeover and giveback work
When takeovers occur
What happens during takeover
What happens after takeover
What happens during giveback
Management of an active/active configuration in normal mode
Monitoring active/active configuration status
Monitoring the hardware-assisted takeover feature
Checking status
Testing the hardware-assisted takeover configuration
Checking hardware-assisted takeover statistics
Description of active/active configuration status messages
Displaying the partner's name
Displaying disk information on an active/active configuration
Enabling and disabling takeover
Enabling and disabling immediate takeover of a panicked partner
Halting a node without takeover
Commands for performing a takeover
Configuration of when takeover occurs
Reasons for takeover
Specifying the time period before takeover
How disk shelf comparison takeover works
Configuring VIFs or interfaces for automatic takeover
Managing an active/active configuration in takeover mode
Determining why takeover occurred
Statistics in takeover mode
Management of emulated nodes
Management exceptions for emulated nodes
Assessing the emulated node from the takeover node
Using the partner command without arguments
Using the partner command with arguments
Assessing the emulated node remotely
Using telnet
Using Remote Shell
Emulated node command exceptions
Performing dumps and restores for a failed node
Giveback operations
Performing a giveback
Replacing failed disks prior to attempting giveback
Initiating normal giveback
Troubleshooting if giveback fails
Forcing giveback
If giveback is interrupted
Configuring giveback
Option for shortening giveback time
Setting giveback delay time for CIFS clients
Option for terminating long-running processes
Setting giveback to terminate long-running processes
Enabling automatic giveback
Configuring network interfaces for active/active configurations
Understanding interfaces in an active/active configuration
Shared, dedicated, and standby interfaces
Interface roles in normal and takeover modes
Takeover configuration with shared interfaces
Takeover configuration with dedicated and standby interfaces
Interface types and configurations
Making nondisruptive changes to the virtual interfaces
Configuring dedicated and standby interfaces
Downloading and running the Cluster Configuration Checker utility
Troubleshooting takeover or giveback failures
Removing an active/active configuration
Ensure uniform disk ownership within disk shelves and loops in the system
Disabling the software
Reconfiguring hardware for stand-alone operation
Management of storage in an active/active configuration
Managing Multipath Storage
What Multipath Storage for active/active configurations is
How the connection types are used
Advantages of Multipath Storage for active/active configurations
Requirements for Multipath Storage (FAS and SA systems)
Requirements for Multipath Storage (V-Series systems)
Determining whether your AT-FCX modules support Multipath Storage
Cabling for Multipath Storage
Adding storage to a Multipath Storage loop
Managing non-Multipath Storage
Overview of adding storage to non-multipath configurations
Adding disk shelves without powering down
Restrictions for addition of disk shelves to an active/active configuration
Adding storage to an existing non-multipath loop
Adding a new non-multipath loop
Adding storage to fabric-attached MetroClusters
Upgrading or replacing modules in an active/active configuration
About the disk shelf modules
Restrictions for changing module types
Best practices for changing module types
Testing the modules
Understanding redundant pathing in active/active configurations
Determining path status for your active/active configuration
Upgrading an LRC module to an ESH or ESH2 module
Hot-swapping a module
Disaster recovery using MetroCluster
Conditions that constitute a disaster
Ways to determine whether a disaster occurred
Failures that do not require disaster recovery
Recovering from a disaster
Restricting access to the disaster site node
Powering off the node
Fencing off the node
Forcing a node into takeover mode
Remounting volumes of the failed node
Recovering LUNs of the failed node
Fixing failures caused by the disaster
Reestablishing the MetroCluster configuration
Rejoining the mirrored aggregates
Re-creating the mirrored aggregates
Nondisruptive hardware changes
Replacing a component nondisruptively
Removing the old hardware
Installing the new hardware
Controller failover and single-points-of-failure
Single-point-of-failure definition
SPOF analysis for active/active configurations
Failover event cause-and-effect table
Feature update record
Abbreviations