You are currently on IBM Systems Media’s archival website. Click here to view our new website.

POWER > Systems Management > High Availability

PowerHA for IBM i Provides High Availability for a Broad Spectrum of Users


Illustration by Otto Steininger

Every business, regardless of size, must be prepared to keep its systems running during adverse conditions such as application or middleware failures, power outages and weather events. In fact, some organizations are required to have systems that will automatically switch to a backup system in an emergency. To get the job done, many have relied on logical replication software at a cost of numerous IT hours each week. But for Power Systems* users, the IBM PowerHA* SystemMirror can free clients from worrying about maintaining a high-availability solution, and it takes just minutes a week to monitor.

The PowerHA solution is designed for small to midsized businesses (SMBs) as well as large corporations. No matter if the backup system is within a few miles or hundreds of miles from the primary system, there’s a PowerHA package that will suit IBM i users.

For many IBM i clients using logical replication as a means of backup, PowerHA SystemMirror for i will be a new concept. PowerHA technology uses shared storage clustering to provide high availability. Shared storage clustering is well known in the UNIX* world but is a relatively new concept for IBM i. The underlying architecture of shared storage clustering is a volume group of disks that can be connected to more than one system. For IBM i, the shared storage is known as an independent auxiliary storage pool (IASP).

“In its most basic form, shared storage clustering is about switching access to the shared disk pool between systems, so if system A goes away, system B takes over the data and production,” says Steven Finnes, Power Systems Product Management, PowerHA. “The PowerHA basic architectural construct isn’t based on replication—although most of our clients replicate the IASP data. PowerHA is about sharing storage between systems.”

With PowerHA technology, the application is writing data continuously to the pool of shared disks and ownership can be transferred at any time. With one command, production work can be switched from the primary node to the secondary node. Users often won’t even know it happened.

 

Finding the Solution for You

IBM offers two PowerHA for i editions: a Standard Edition for data centers or single sites, and an Enterprise Edition, which is a multisite solution supporting up to three sites. Both editions use hardware replication methods to keep the systems in the cluster in sync. Data is continuously replicated between systems in the cluster using external storage or internal storage. If the systems are located within approximately 25 miles of each other, the replication can be synchronous. If the systems are hundreds of miles away, then an asynchronous replication method can be used.

Geographic Mirroring (geomirror) is the host-based replication that comes with the operating system and is typically used with internal disk. With hardware replication methods, synchronous writes to the remote disk are synchronous to the application state; the application won’t proceed to the next step until the write has been completed remotely and locally. Clients will need to use asynchronous replication if they require geographically dispersed locations but can’t tolerate application performance delays. “It gives you high availability with geographic dispersion,” Finnes says.

IBM i clients running on a Power Systems 720 server with up to 5 TB of internal storage may find geomirror to be a good choice for their PowerHA deployment.

Organizations using IBM i with a storage area network (SAN) such as IBM System Storage* DS8000* or IBM Storwize* V7000, have two storage-based replication options. Where the backup system is within a metro area, Metro Mirror would be the right selection. For asynchronous backup over long distances, the choice is Global Mirror. With both Metro Mirror and Global Mirror, the volume group (IASP) is being replicated in real time.

Clients using SANs will find that in addition to PowerHA real-time replication, a capability called FlashCopy* lets them instantly create a flash copy of the production data in a partition they can subsequently save to tape at their leisure. This technology effectively eliminates the common backup window problem for clients. The PowerHA solution integrates the Storage Server copy services technology, which includes FlashCopy, Metro Mirror and Global Mirror, with the IBM i clustering technology into a single interface for automation of all key aspects required for availability/recovery management. The advanced copy services option from IBM Lab Services extends the base PowerHA functionality to three sites and can provide an even greater level of automation and customization.

“We have a solution for any IBM i client,” says Jenny Dervin, technology leader, PowerHA for i. “Geomirror is perfect for smaller clients with internal storage. As clients move into external storage technologies, we have multiple technologies available. You can combine those into customized PowerHA solutions.”

 

The i-Hosting-i Advantage

SMB clients with internal storage can take advantage of a unique PowerHA implementation utilizing i-hosting-i technology. In this implementation, the OS runs as a partition hosted by another partition. The OS is placed into a network server storage space, which is created in an IASP. The entire system is replicated to the backup in real time. The IBM i system is the hardware layer and i-hosting-i exists as a virtual layer inside the hardware. Figure 1 illustrates the difference between traditional geographic mirroring and i-on-i geographic mirroring.

The i-hosting-i PowerHA implementation is well suited for smaller clients because of its simplicity, according to Douglas L. Bidwell, systems engineering manager at DLB Associates Inc. Bidwell has performed several i-hosting-i implementations for SMB clients and says the PowerHA solution can be up and running in one day using i-hosting-i. “A traditional IASP implementation is more involved,” he explains.

“With i-hosting-i implementations, no IASP application adaptation is required,” says Bidwell, who recently helped a client with 23 third-party packages implementing an i-hosting-i PowerHA solution. “There’s nothing for IT to do in the sense of splitting out the database into the IASP. The entire operating system and data are replicated in real time.”

Bidwell compares using PowerHA to traveling with a single suitcase. “If I take one suitcase on the plane with me, I’ve got it all,” he explains. “The more I can put on an IASP, the easier my travels are going to be. With logical replication, everything requires a different piece of baggage. With some third-party applications, you’ve got thousands of pieces of baggage, making it horrible to administer. With i-hosting-i and PowerHA, the whole system, not just the IASP, is in the suitcase, so I can grab it and go.”

 

Shirley S. Savage is a Maine-based freelance writer. Shirley can be reached at savage.shirley@comcast.net.



Advertisement

Advertisement

2019 Solutions Edition

A Comprehensive Online Buyer's Guide to Solutions, Services and Education.

POWER > SYSTEMS MANAGEMENT > HIGH AVAILABILITY

Investments Offset the Costs of Data Loss

Data is Money

A recent survey explores the state of Power Systems resilience

POWER > SYSTEMS MANAGEMENT > HIGH AVAILABILITY

Data Backup Options Balance Risk and Cost

IBM Systems Magazine Subscribe Box Read Now Link Subscribe Now Link iPad App Google Play Store