Scenario: Global Mirror
This scenario describes an PowerHA® solution that is
based on external storage and provides disaster recovery and high
availability for storage systems that are separated by great distances.
Global Mirror is an IBM® Systems
Storage solution that copies data asynchronously from the storage
unit at the production site to the storage unit at the backup site.
In this way, data remains consistent at the backup site.
Overview
The Global Mirror solution provides
a disaster recovery solution by using external storage units across
long distances. The independent disk pool is replicated between the
external storage devices to provide availability for both planned
and unplanned outages.
This scenario covers the configuration of IBM-supplied IBM i high availability technology
and does not provide installation or configuration instructions regarding IBM
System Storage®
DS8000® series. This information assumes that an IBM
System Storage solution is already in place before
the IBM i high availability configuration.
For installation and configuration information about DS8000, see Copy Services in the IBM
System Storage
DS8000 Information Center
.
A similar scenario can also be accomplished with other
system storage technologies. For more information on storage technologies
that support Global Mirror, see PowerHA supported storage servers.
Objectives
Cross-site mirroring with Global Mirror provides the following advantages:
- Replication is entirely managed by the external storage unit, thus no IBM i CPU is used. Replication continues in the storage unit even when the system experiences a system-level failure.
- Availability for business resources during planned or unplanned outages, which include maintenance outages or software/PTF related outages as well as disaster recovery.
- Fast recovery times when used with journaling. Journaling recovers data more quickly in the event of an unplanned outage or failover. Journaling forces data changes to disk where the mirroring is occurring. If you do not use journaling, you might lose data that is in memory. Journaling provides recovery of these data-level transactions and helps with the recovery times.
- The ability to use the FlashCopy® function on the source or target side of Global Mirror.
- The solution requires IBM System Storage DS8000 server hardware.
- To achieve acceptable performance, consider using redundant communication paths and adequate bandwidth.
- There is no concurrent access to the disk pool.
- A consistency group is required for the Global Mirror target copy. A consistency group is not required for the Global Mirror source copy, but it is highly recommended.
- Reverse replication occurs automatically on a switchover only if the new target has a consistency group. Reverse replication never occurs automatically on a failover.
- When reverse replication does not occur on a switchover
or failover, the configuration will consist of two source copies.
- If the desired target copy node has a consistency group, then a reattach operation will convert it to a target copy and automatically initiate replication.
- If the desired target copy node does not have a consistency group, then recovery requires manual intervention with the System Storage DS8000 Storage Manager interface to initiate replication and synchronize the current source and target.
Details
The following graphic illustrates this solution:
Configuration steps
Complete the following steps to configure the high
availability technologies that are associated with this scenario using
the PowerHA graphical
interface: - Complete planning checklist for cluster
- Create a cluster
- Create a cluster administrative domain
- Start a cluster administrative domain
- Create an independent disk pool
- Add monitored resource entries
- Make the independent ASP highly available
- Configure Global Mirror
- Vary on the independent ASP
- Perform a switchover to test your high-availability solution
Complete the following steps to configure
the high availability technologies associated with this scenario using
other interfaces:- Complete planning checklist for cluster
- Create a cluster
- Add a nodes
- Start a node
- Add nodes to a device domain
- Create cluster administrative domain
- Start a cluster administrative domain
- Create an independent disk pool using Configure Device ASP.
- Add ASP copy description
- Add monitored resource entries
- Create a device CRG
- Start a device CRG
- Start ASP session
- Configure global mirror session
- Make disk pool available
- Perform a switchover to test your high-availability solution