Questions to Ask Yourself When It Comes to Backup and Recovery

steve_pitcher
Steve Pitcher

When people talk to us about backup and recovery, one of the most common questions is “how often should I be doing a full system save?”

Given we’re about to install PTFs or upgrade the OS, my response is always, “when was your last full system save?” What I do next is show them the QSAVSYS data area to at least find out exactly when the last SAVSYS (Save System, or Go Save option 22) was done. Now, this isn’t proof positive that a Go Save option 21 was taken because a SAVSYS is a part of a backup using Go Save option 21. But in terms of a recovery, a SAVSYS will allow you to at least help get your Licensed Internal Code and operating system going from a bare metal recovery.

Check your last SAVSYS by reviewing the QSAVSYS data area with the following command, then review the Save Date/Time value of the object.

DSPOBJD QSAVSYS *DTAARA DETAIL(*FULL)

The question of frequency is still not answered.

It depends. Personally, I like to see one done at the very minimum once a year. If you can get away with once a quarter that’s better. We know shops that do it once a week and even a few who do an automated version of Go Save option 21 every single night. A business’ tolerance for downtime will often dictate frequency. Some have the luxury of being able to bring their system down every night. Many others do not.

Do you know when you last did a full system save?

If you’re not sure, then it’s probably been a while. The above command will give you a good idea when your last full save was.

Was it on physical tape? If so, where is it? 

Was it on a virtual tape library? If so, do you know what volume it was written on?

How often do you update your PTFs?

If your Licensed Internal Code, operating system, and IBM licensed programs are changing every quarter due to PTF updates then doing a full system save or at least a SAVSYS either before, after or both isn’t a bad idea. Let’s say last did a SAVSYS in August and you’ve updated your PTFs four times since then. If you had a total system failure and had to recover, all those PTF updates would NOT be recoverable.

A good rule of thumb would be to ensure you have at least a SAVSYS before any major system changes like an OS upgrade or a PTF update and that your tape media is easily accessible and identifiable. If your last full system save was from two years ago, your recovery media is available in a jiffy and you haven’t changed your system since…then that should suffice. However, if you’re relying on media that you think has been sitting there for two years you run the risk of it being overwritten without you knowing it. Your media is only as good as your last recovery.

Have you ever tested that old save media?

I’m sure my mom thinks that an old VHS tape of my 5th birthday party still has my birthday party on it. Chances are, it’s been either fully or partially overwritten sometime since 1984. If by some miracle it’s been left untouched, the tape has certainly degraded since then. While LTO lasts much longer, the point is that time allows capacity for change and we should  not get too comfortable in assuming what’s on media is what we think.

Case in point, we had a customer have a full system failure back in the fall.

They sent us two tapes for recovery: a full system save (from 2012) and a nightly backup from two days before the crash. The full system save tape did not have the LIC or OS on it. It had a total of six libraries. The nightly backup gave us user libraries, user profiles, configuration and DLOs. No IFS. We built them a partition with LIC and OS (not their LIC and OS) and got them back up and running with their nightly data tape on iInTheCloud.com. They were extremely lucky.

Hopefully these questions should help you determine what’s best for your systems and your business. One thing is for certain — nobody has ever been at fault over having too many full system saves.

Looking for more? Contact us and we’ll help you out.

Tagged with: , , , , ,

Leave a Reply