TSM and IBM i: Not What It Could Be

December 5, 2011 Alex Woodie

It is not unusual in the IBM i world to discover that IBM as a whole pays more attention and spends more time developing solutions and utilities that run on and support third-party operating systems rather than its own IBM i OS. Call it the nature of i beast. But when it comes to IBM’s strategic backup and recovery product, Tivoli Storage Manager (TSM), at least one IBM business partner wonders whether it’s time that IBM i becomes a fully supported TSM client.

For years, STORServer has been in the business of packaging, selling, and supporting backup appliances based on TSM software and IBM X86 servers. The Colorado Springs, Colorado, company is an adamant supporter of IBM and, in particular, TSM, which it regards as the most feature-rich backup and recovery product in the market–but also its most misunderstood.

The problem with TSM–complexity–is also its strength, according to STORserver chairman John Pearring. “It’s complicated. But with complication comes robustness,” he says. “TSM is written basically as a command line product. IBM has a GUI console that works for the extremely large enterprise shops. It does not work well or translate well to a midmarket customer. That is our space.”

STORserver’s GUI management console eliminates the need to manage TSM from the command line, Pearring says. “Everything that they need to do with Tivoli Storage Manager is in our console. It takes the difficulty out of TSM completely out of the box,” he says. “And the fact that we provide an appliance, the other difficulty of TSM, which is to implement and configure properly, we take care of that also.”

With a large swath of IBM midrange customers using its disk-to-disk (D2D) backup solutions dating back to the late 1980s, it’s no wonder that STORserver has run across its fair share of AS/400, iSeries, i5, System i, and now IBM i-based Power Systems customers. Even with a target market of midrange customers protecting databases of 100 TB and under, you would think there would be more than a few IBM i shops willing to consolidate their backup procedures with TSM.

However, few of them have ever chosen to use TSM to run backups for the platform. Instead nearly all of them choose to continue running BRMS to back up the IBM i servers. This leaves the customers with two separate backup systems: BRMS for IBM i, and TSM for everything else.

Now, IBM does offer some support for backing up IBM i data with TSM, and this is documented well at on IBM’s BRMS site. But this is by no means an enterprise backup and recovery solution, and there are so many limitations that it makes one wonder: Why would any IBM i shop choose this?

For starters, IBM stresses that TSM for IBM i is only for “low-volume” backups of distributed IBM i systems; it won’t work well in large production settings and it can’t be used for standalone systems. Secondly, since you cannot save IBM i system objects to the TSM Server using the TSM for IBM i software, you cannot perform a full IBM i system recovery with TSM. Saves need to be duplicated to tape after they’re saved to the TSM servers, and some save-while-active functions are not supported. Also, you can’t schedule any IBM i backup activities from the TSM server; backups can only be set from the IBM i job scheduler.

The BRMS Application Client is the way that IBM i hooks into TSM, but it’s not a full TSM client in the same sense that IBM offers TSM clients that provide backup/recovery and data protection capabilities for Windows, Linux, Unix, and other operating systems. The BRMS Application Client, which uses the TSM APIs, basically turns a TSM server into “just another device that BRMS uses for your save and restore operations.” While the BRMS Application Client for TSM may help speed up your backups at offsite locations and minimize the handling of media there, you still need to have a tape drive on hand for local restores, and it doesn’t get you off the hook for using those tape drives regularly. And you’re still totally reliant on BRMS.

Pearring would love for IBM to change this, and to create a full TSM client that runs on the IBM i server, and turns the IBM i server into a full-fledged member of Team TSM.

“In six months, IBM could make an announcement. They could move very quickly on this and we urge them to do so,” Pearring tells IT Jungle. “They have a lot on their plate. We understand they have their priorities. But the AS/400 group within IBM should make that change, absolutely. We would be so happy to roll it out. And I know our customers would love to see a roll out of the latest version of TSM and these new capabilities that we could be connecting up to the AS/400, but it’s on IBM’s plate where that has to be done.”

Pearring admits he doesn’t know why the change hasn’t been made. “I don’t understand the intricacies. I don’t know if it’s political. I doubt it,” he says. “I think there’s probably some hardware issues going on there. But whatever’s going on, I would urge them to go ahead and get that fixed. There are other people outside of IBM encroaching into the iSeries data protection field, and it’s not necessary. IBM can do that work.”

The use of LTO tape emulation is the quickest way that third-party storage vendors can get access to IBM i. In the last few years, a host of vendors of D2D and virtual tape library (VTL) systems have come into the IBM i space. But emulation is not the ideal way to go, Pearring says.

“They’re not fully supporting–they’re emulating. That’s a word,” he says. “They’re hanging storage. They’re not fully supporting iSeries. IBM is the one who can fully support it. Everybody else is literally doing a kludge of some sort to attract that iSeries business. You can emulate the tape and hang storage, you can do those things, but that’s not what those iSeries customers are looking for,” he continues. “They’re looking for a full implementation of TSM and iSeries, so that it just fits into the rest of their operations. They’ve done it with the hardware, the servers, they’ve done it with the storage. It’s time to do it with data protection.”