After a week beyond the end of End-to-End Root Cause Analysis, I have the podcast link to share. Near the end of the blog are a series of screen shots, similar to the preceding days.
With the now customary final class day being a half-day, the material left to the end was the last 50 or so pages in the book, along with a recap of the day before. We primarily looked at various tracing tools and techniques. One interesting tidbit was the revelation of “transactional correctness, a business logic verification process that purportedly can discover bugs such as user exits with frowned-upon commit phases (or out-of-phase commits perhaps). The book example was somewhat rigged for demonstration purposes; the more useful discovery was the trace panel button that can trigger this audit.
When I looked in SDN for the string “transactional correctness” I only found a few articles, and no forum messages that appear to discuss this topic directly. Either I’m not searching well, or this is a well-hidden gem. I found it in the SAP GUI for our 4.7 systems, so it seems to have been there for some time.
A bit of our tracing attempts were thwarted by more than one of us creating and viewing the same trace file on the same system. If that happens in your shop, you need to coordinate with any other investigators. It doesn’t make sense to duplicate the effort.
A few of the SAP Press books I have purchased have been helpful (particularly Schneider’s Performance Optimization Guide book – so good that my colleagues have borrowed, long term, my copies of the first and third editions respectively. I think it is time for a new edition, and a new Amazon review …)
Books like “Performing End-to-End Root Cause Analysis Using SAP Solution Manager” are unappealing to me, based on the prices to page count ratio. I’ve made the mistake of buying books with under 100 pages before, and it’s a problem. I understand the economics of such “small market” editions – why can’t we get a PDF for a fraction of the printing cost?
Anyway, back off the soap box. I bought this book:
Solution Manager Enterprise Edition – SAP Press
As Gertrude Stein said about Oakland California, “not a lot of there there.” Reading an overview of features that took us days to dig into, using a 500+ page manual, covered in under 40 pages, leaves me wondering why I bothered to order the book.
In Chapter 8, Section 8, the above book covers “End-to-End Root Cause Analysis.” Upon reading that section at the conclusion of the course, I’m left with that “buyer’s remorse” feeling that one gets when the whole just isn’t much. Phrases like “identify the affected component while minimizing the time needed to solve the problem,” and (this one is my favorite) “diagnosis [sic] agents installed on every managed system … preconfigured and shipped by SAP.” That latter sentence just makes me gasp with the realities omitted from the conversation. If you don’t understand why I disagree, you don’t get root cause analysis of a complex system environment.
Though I am not the teacher, I’m the reviewer, so I’ve turned to the classic of product reviews, Consumer Reports, for the style of grading, known as “Harvey Balls.” Through wikipedia, I found images of the color codes in a sharable format (under “Creative Commons Attribution-ShareAlike 3.0 License” or “GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.”) I’m unsure how to attach either of these to the derived work that follows, but as they’re simply circles with colors in full or half, I think we’re on solid ground here.
Here’s the scale, from left to right, excellent to poor. If you’re color blind in anyway, the left disk has a small hole in the center, and it’s red.
What I wanted to do was rate the course content in two ways – how well it might work, and how well I think it will work. And when I say “course content” I mean applying the lessons in the class to the versions of Solution Manager and other software in a business environment. Let’s start at the top level, focusing on the primary tool choices, or subject areas of Root Cause Analysis. I might use different phrases for what I do when looking at application or system performance instead of Workload Analysis, but that’s what SAP calls it.
To put the following charts another way: “fit and finish” is how well the product behaves; “applicability” is how much work I expect to have to do to get any benefit from it.
|Function||Fit + Finish||Applicability|
I planned to dive deeper, but as time flies by, I’m going to review one component here, and perhaps discuss others in the blog comment threads.
For “Workload Analysis”, I’m perplexed by the overlap with Central Performance History. Both seem to be extensions of ST03 metrics, with an implied agent infrastructure and a new (additional) repository, probably BW. That’s great, if it simply comes out of the gate working. But as far as I’ve seen, it doesn’t. The charts that we saw in class were okay, with the caveat that they didn’t always work. In the real world, when I looked at a Solution Manager 7, EP1, system, there was not much similarity. I think the concept is grand (web based, graphic views, leaning in the dashboard direction); I’m concerned there isn’t a wiring diagram, and with shorts or blown fuses, putting the pieces back together is reminiscent of the Catch-22 stove repair saga.
Jose kindly agreed to a podcast, and here is the link:
A brief recap:
|2:00||Moving parts – discrete components|
|2:38||Practicalities – central hub|
|4:30||Web / Java / ABAP|
|5:40||Troubleshooting / Proactive|
|7:00||Six month checkup?|
|8:30||Virtual training environment|
|9:30||Wishlist future tools|
|10:30||No camera / remote control sharing|
|11:30||Social networking vector for students|
|13:00||SCN exposure; blogging first|
|14:30||Chef and critic|
|19:00||The reporter; urban dictionary|
|20:45||Green room, publication feedback|
The way I approached this conversation was as if I chatted with the instructor on break or after class. We hit a few wish list items, not just for the sustainability factors of virtual training, but evolution of the SAP community. It’s not Jose’s responsibility, but I’d think more students would sign up and attend if the virtual class was discounted heavily over the physical class. The instructor makes out better, less travel expenses, and students should learn the material as well. Potential wins all around.
The screen shots, mainly from class:
We get to rate the instructor:
My dilemma uploading the podcast. This hung every time for several days, including the certificate error on this screen.
Video (for break time)
The virtual matriculation document
What shall I take next?
Funny I should ask, as an auto-email said I should take E2E200, or maybe it was E2E300. It doesn’t matter, as I’m not going to get better educated until I get the pieces to work that I’ve inherited so far.