Admin Alert: The Great CBU Survey and More
Published: October 7, 2009
by Joe Hertvik
Let's change it up a bit this week. First, I'm initiating the Admin Alert Great CBU Survey where I'm asking to collect information from users and vendors who work with Capacity BackUp (CBU) systems. Then I'll follow-up on and add a little more information to some previous articles I posted. That's a lot of information to cover in 1,200 words or less, so let's get started.
Chapter 1: The Great CBU Survey
After reading Part 1 and Part 2 of my columns on performing live CBU switch tests on a System i box, reader Don K emailed me the following comment:
I'd like to see a poll showing how many users perform switch tests, as well as how frequently and how often they perform the tests. At my previous employer, we did quarterly tests and ran for a week switched over. My current company switches over on a Saturday morning and only leaves it switched for four or five hours, which doesn't leave much time to test all functions. This is frustrating because you have to hurry up to allow the users back on while you're trying to solve CBU problems.
This reminded me of a recent conversation I had with a vendor. He seemed to think that a large number of i5/OS shops with CBUs never perform switch tests and only a small percentage of shops ever perform a live switch over.
That got me thinking about how CBUs are actually used in i5/OS shops. Are they a vital, tested infrastructure component or are they sitting around with a configuration that isn't quite ready to go? Not that I'm suspicious, but I'd like some confirmation as to how many shops are actually taking advantage of this technology.
All of which leads me to the great CBU survey, an informal polling to see what i5/OS CBU owners are doing with their high availability machines. If you own a CBU and corresponding replication software, please drop me a line via the IT Jungle contact Web page to let me know what you're doing with it.
If enough readers provide me with the following information, I'll be glad to collate it and share it in a future column as a quasi-state of the i5/OS CBU statement.
Here's what I'm looking for:
- What is your name? (Names will be kept confidential.)
- How long have you owned your CBU?
- Is your CBU currently replicating a production partition?
- Have you ever switched production processing over to the CBU?
- How often do you perform switch over exercises? How long do you stay switched over?
- Has your applications team validated that production applications work on the CBU?
- Have your users ever run production processing on the CBU?
- Have you ever run a normal day's business processing on the CBU?
I'm also looking for high availability software vendors to participate, too. Don't just tell me what's in your FAQ. Give me an idea of what your customers are doing with the technology.
To make things sweeter, all entries will win an Admin Alert No-Prize. It's similar to a Marvel Comics No-Prize, but it's specific to Admin Alert content.
Chapter 2: Help for Ray and His Client Backup Issues
In my September 23 column, I introduced Ray, a consultant who had trouble convincing his client to perform regular reliable system backups. His client preferred to only backup his user libraries on a daily basis, and the client didn't even want to backup the QGPL or QUSRSYS libraries. With this lackadaisical backup policy, Ray feared that the client was heading for a disaster. I asked my readers for advice on Ray's problem.
Reader Roger Paul had this to say:
I wonder if Ray could simulate a disaster for his client? Ray and his client could take the backup media to a disaster recovery site, load it on a recovery machine, and then maybe the client would see whether his backup technique really works.
While this is a good idea, I'm not sure it goes far enough. The restore should be accompanied by a disaster recovery drill, where users test the restored machine to see if they can process data. Having users run actual production data might demonstrate how the client's backup process performs in a pinch.
Reader Peter Barnum wrote in with this observation:
Another gotcha is the innocuousness of creating user libraries where the library name is equal to the user name. For example, we use the QSYSOPR profile for normal operations and save query data to a library that is also named QSYSOPR. Even though I save all user libraries every week, I never realized that the QSYSOPR library was never being saved. This was because when the Save Library (SAVLIB) command performs an all user libraries backup (*ALLUSR), it skips saving any libraries that start with the letter 'Q'.
It's always the things you don't know that kill backup effectiveness. Application data can migrate to the darnedest places and if you don't backup everything with your backup strategy, the things you miss can cause a lot of problems.
Chapter 3: Viewing Status History Information in PC5250
Last week, I discussed how some users aren't able to start new sessions using PC5250's run the same function. Instead, they receive a status bar message that reads "Disconnected." While I demonstrated how to fix the problem, I didn't mention that you can troubleshoot the issue by using PC5250's Status Bar History feature.
PC5250 displays error messages on line 25 of its terminal display. This is dependent on turning on the 5250 Line Checkbox in the Windows Setup panel (which is accessed by selecting Edit→Preferences→Appearance→Setup from the PC5250 menu bar). When checked, PC5250 displays the most recent error message on line 25 of the display. However, there are two problems with how PC5250 displays status bar messages.
- You can only see the most recent error message.
- It doesn't display other messages that aren't error messages.
To see all the messages that PC5250 has generated, use the Status Bar History screen. Press View→Status Bar History from the PC5250 menu bar and you'll get the following screen that displays the most recent status bar messages for this session.
So when troubleshooting PC5250 connection problem, you can use this feature to display the session's most recent error and status messages.
The Cost of Not Backing Up
The Road to Live CBU Fail Over, Part 1
The Road to Live CBU Fail Over, Part 2
When PC5250 Run the Same Doesn't Run
Post this story to del.icio.us
Post this story to Digg
Post this story to Slashdot