Admin Alert: Four Ways To Move An IBM i Partition, Part 3
January 29, 2014 Joe Hertvik
The last two issues, I’ve discussed different ways to move an IBM i partition from one location to another. I talked about using traditional backup and restore techniques, as well as using logical replication High Availability (HA) software for partition migration. This week, let’s finish the series by looking at using externally connected storage including SANs, PowerHA SystemMirror for IBM, and Live Partition Mobility for moving partitions.
The Four Ways To Move An IBM i Partition
To recap, there are four basic ways to move an IBM i partition.
Externally Connected Storage
When you configure a Power i machine (IBM Power systems box running IBM i partitions) with externally connected storage, it becomes much easier to migrate your partitions from one Power server to another.
Let’s say you purchased a Power 6 machine with externally connected storage in 2010 and now it’s time to upgrade to Power 7+. With internal storage, the system and all its disk drives are basically one and the same unit. This means you have to shut down and perform an in-place upgrade to install new hardware for your partition.
When you use externally connected storage such as a SAN, the system unit (including its non-disk resources) and its disk subsystems are two independent components working together. The system unit contains the processors, memory, Ethernet cards, and other resources while the external storage unit only provides the storage. The external storage unit could even be shared with other servers, so it might not even be exclusively connected to your Power i setup.
When system unit and disk are separate components, you can swap or upgrade each of the components independently of the other. So if I wanted to swap out my Power 6 box for a new Power 7+ machine, I could install the new Power 7+ unit, disconnect the disk containing my Power 6 partitions and data from the old Power 6 machine, and reconnect them to the new Power 7+ machine. Migration accomplished without a lot of hassle. Or if I wanted to swap out the external disk, I could do so without upgrading the system unit. Using external disk provides more flexibility with upgrades than using internal disk. You can even put your system unit and disk on separate upgrade and leasing schedules, so you can upgrade them at different times.
Why Flash Copy Isn’t Suitable For Partition Movement
Flash copy is a technology available on many SAN storages devices. With flash copy, you can take almost instantaneous copies of your IBM i partition (with little disk usage) and then back up the copied partition data, leaving your production partition free to continue processing without interruption.
Flash copy is great for eliminating backup windows, but it has limited use in moving a partition from one system to another. Partitions cannot be flashed between different SANs so flash copy is limited to copying partitions on the same SAN you are currently running on. So while you may be able to use flash copy to create a backup copy of the partition you want to move, you will still have to use one of the other techniques to actually move your source partition to another machine.
PowerHA SystemMirror For IBM i
With IBM PowerHA SystemMirror for IBM i (PowerHA SystemMirror), there are several ways you can use clustering technology to move an IBM i partition from one machine to another. PowerHA SystemMirror uses shared storage clusters and storage-based replication rather than logical replication to move data and applications between systems. (I discussed using logical replication solutions for moving data in Part 2.)
Using IBM’s PowerHA System Mirror for IBM i package, your applications and data are deployed into an Independent Auxiliary Storage pool (iASP). iASPs can be created on internal or external hard drives. iASP data is kept current between iASPs by using either LUN-level switching, host-based mirroring, or storage server-based mirroring such as you’d find on an IBM SAN. You can use geographic mirroring (geomirroring) to keep your iASPs in sync if you are running IBM i 7.1. If you are using an IBM storage device, you can keep your iASPs in sync using either of two IBM storage-based mirroring solutions: Metro Mirror for synchronous mirroring over distances of around 25 miles or less; or Global Mirror for asynchronous mirroring over distances greater than 25 miles.
It’s important to note that PowerHA SystemMirror is an extension of IBM’s storage management architecture. For storage-based mirroring, it requires IBM Storage Area Network devices such as the DS8000 series or the IBM Storwize V7000.
Like the logical replication techniques I discussed last issue, PowerHA SystemMirror is primarily intended for high availability and disaster recovery situations (solutions that involve moving IBM partition processing from one machine to another). When set up correctly, PowerHA SystemMirror can be used to switch processing between a remote and local location so processing can continue at one location while the Power system disk at the other location is being upgraded.
When I’ve talked to people about PowerHA, they love using it for clustering, HA, and DR but the biggest issue is set up. If you’re clustering over any kind of distance, you’ll have to set up iASPs for disk storage at all of your clustered locations. Going from IBM i non-iASP storage to iASP storage can be a major undertaking. It may require you to segment different components on your IBM i hard drives and it may even cause you to reload different applications and components onto different iASPs or ASPs. You have to ensure that all of your applications are cluster-proven and able to work in this environment. In general, the advantages of using PowerHA SystemMirror for high availability and disaster recovery are great but there could be significant work getting it set up in your environment: work that might best be accomplished in conjunction with (ironically) a Power systems disk upgrade.
Consult IBM and your business partner if you want to consider installing PowerHA for SystemMirror as a clustering, high availability, and disaster recovery solution on your system.
Live Partition Mobility
Live Partition Mobility (LPM) is a relatively new component of the PowerVM Enterprise Edition hardware feature. It allows you to move a running server from one Power system to another.
Controlled through the Hardware Management Console (HMC), LPM requires virtualized storage and two relatively equivalent Power systems (same CPU, memory, interactive features, P-groups, etc.). LPM allows you to transfer an active or an inactive partition from one machine to another. Live Partition Mobility works by doing two things:
By doing this, a partition can quickly be moved between systems with little to no disruption to running jobs. Existing jobs are suspended and restarted as they change systems.
Because the migrated system disk is moved (pointed) from the source partition to the target, LPM is similar to the externally connected disk solutions I discussed above. The difference is that an LPM transfer is almost instantaneous and unnoticeable to the users or your jobs.
Also because LPM relies on pointing disk between machines, LPM seems to work best in a local environment where the receiving machine is nearby to take advantage of faster system to disk connections. So it doesn’t work for migrating a partition across a remote network.
The End . . . For Now
This concludes my tour of available technologies for moving an IBM i partition from one system to another. Please feel free to email me with any questions at firstname.lastname@example.org.
Follow Me On My Blog, On Twitter, And On LinkedIn
Check out my blog at joehertvik.com, where I focus on computer administration and news (especially IBM i); vendor, marketing, and tech writing news and materials; and whatever else I come across.
Joe Hertvik is the owner of Hertvik Business Services, a service company that provides written marketing content and presentation services for the computer industry, including white papers, case studies, and other marketing material. Email Joe for a free quote for any upcoming projects. He also runs a data center for two companies outside Chicago. Joe is a contributing editor for IT Jungle and has written the Admin Alert column since 2002.