Legacy Environments: What Can Be Done?
What would you do if you had to manage an environment with old POWER machines running AIX 5.3?
By Rob McNelly01/16/2018
I still see it every now and then. For instance, recently I was talking to an executive in a manufacturing organization. This organization is filled with old equipment, production-related machines that cost $100,000 or more new. The makers of that equipment are long gone, but he knows how to maintain it and everything is paid for, so this guy is thrilled. As for his IT gear, he wouldn't care if Windows* XP or DOS was still in place. In his mind he's printing money with these ancient machines. Even though his OS is likely insecure and his hardware could stop running at any point, he doesn't care. IT is simply not part of the equation. He'll run these systems into the ground.
As I said, this sort of thing isn't common, but it's also far from exceptional. You've probably seen it yourself. Typically these operations have standalone systems that sit in a corner and run critical applications from internal disks. Best case, there may be a tape drive and old mksysb backup scripts that are happily running out of cron. Still, do you ever wonder how often, if ever, the tape drives are cleaned, or if new tapes are purchased? Has anyone ever tried restoring from these tapes?
Of course all these systems have been so reliable that no one even checks on them--not that anyone on staff would know about them anyway. The IT guy who took care of this stuff when it was new probably left years ago, and was never replaced.
They say you can't help people who aren't willing to help themselves, but when it comes to these customers, I still want to try. So how would you deal with these types of operations? What are some ways forward to at least try to minimize risk?
I would start with the used market and search for the same model hardware with the same tape drive--although even this option is becoming a challenge. If I could find similar hardware, I'd take a mksysb from the source system and try to restore it onto the "new" box. What I like about a secondary system is that this testing can go on without affecting anything. Plus, if you need to back out, it's as simple as going back to the original machine. At least this provides some peace of mind, because if it all works, you know that there's a good backup and hardware to restore it to.
If that cloned hardware can run, then it's possible to try to upgrade the "new" machine, again with no downtime. Of course this is no sure thing, but surprisingly, some applications will run on a newer version of the OS. If the OS upgrade doesn't work, or it's already apparent that the application won't run on a newer OS version, the next step would be to get AIX to the latest TL version possible. That would at least allow these folks to consider versioned WPARs or try to restore the mksysb onto a newer generation of hardware running in older processor modes.
If I couldn't locate a whole new machine, I'd settle for ordering replacement internal disks and planning for an outage. I'd remove the original disks, install the new ones, restore the OS and try an upgrade. If I had to back out, I could just reinstall the original disks. Of course this method comes with its own risks.
These are all what I'd call "the best you can do" options. None of these solutions are ideal, and none change the grim reality that nearly of all of these customers are without hardware support, OS support or application support. And for whatever reason, this is considered an acceptable risk. If it ain't broke, don't fix it, right?
As I said though, I'm trying to help. Perhaps you can help, too. Hit the comments and tell me what you'd do in this sort of situation.
Rob McNelly is a senior AIX solutions architect doing pre-sales and post-sales support for IBM Premier Business Partner Meridian IT Inc. More →