A blog entry in two parts

 

Yesterday was very frustrating. The roofing guy was late, and did not give an immediate estimate. While this is actually a good sign as I feel they will provide a much firmer estimate than all others, it is delaying the entire decision process. What really got to me, however, was that Pashy's new laptop appears to be damaged on arrival. I understand that it's not my fault, and that Zareason has procedures to handle this sort of thing, but it hasn't prevented me from beating myself up about it. She's already waited long enough for a new system, and now we have to wait even longer for them to sort what in the heck is wrong with the system they sent us. It's difficult not to think that we would have been better off just buying a refurbished Thinkpad from Microcenter...

Work hasn't been any less frustrating lately either. 

My instructor for the cloud course is still MIA. I haven't heard from him at all, and the custom course is to be taught next week. At this point, I don't even know if that class is going to happen now. The main ILT is likewise in murky territory. The problem is that they introduced a new feature in 3.1 called "Quickstart". QS streamlines much of the processes of on-boarding new virtual data center providers, tenants, users, and basic service creation. On one hand, QS is an amazing new feature that greatly simplifies the process of setting up a cloud instance. On the other, it hides a lot of details and complexities that would otherwise be necessary as a cloud deployment matures.

I spent a lot of time yesterday trying to see the difference between QS-created objects, and objects created through the original, manual processes. Service creation is obviously very, very limited. You are stuck with the templates available from the VDC provider. You cannot create multiple deployment definitions. I'm fairly confident that you cannot use QS to create multi-tier services either. It's great if you just want to get started, but you can quickly grow past its capabilities. On-boarding tenants, however, is much more of a serious question.

The manual process for on-boarding tenants involves collecting and entering a great deal of information that simply seems unnecessary in a cloud context. Region? Site? Department? Who cares? All that's really important in a cloud context is who is requesting the resource, and what tenant they belong to. The tenant is the basis of allocation monitoring and billing. Nothing else matters. 


And apparently, it doesn't matter all that much really. My boss told me to just go on vacation for the next two weeks.

I usually don't take nearly that long. Usually I tend to keep my vacation time at a week at most until the end of the year. Then I usually take everything at once. Maybe I'm getting too old for that sort of thing, and I need to take more time off just to retain my sanity. The problem is that I immediately start to think about the things I could or should do in that time. A few ideas do come to mind, including playing with the Raspberry Pi, working on the basement, and of course, Drupal. The first two require something that's always lacking -- money.

Admittedly, the Raspi would require less money than the basement. If I wished, I could pick up a new Pi at Microcenter, or order one on Adafruit. Perhaps, however, I should set down a specific purpose before buying new hardware. I have wanted to repurpose the existing Pi I have as a wireless access point. It wouldn't be hard to do, but it would require a day or so to set up. A trip to Microcenter might not be a bad idea either way, as it would allow me to hook it up with a monitor so I'm not taking over the television. I can get that at Microcenter tomorrow.

The basement is another issue, really. I would rather have the windows replaced with block *first* before I start to do anything else. That task will be contracted out, but it's one I have to decide on soon otherwise wait all winter for it to happen. I suppose I could set up a contracting appointment with Home Depot and see how much that'll run me. 

Drupal stuff is another issue. Like always, it's best if I try to focus on one or two tasks, rather than vague goals. I have been making considerable headway in trying to understand how Drupal 8 works. I'm about to the point of playing with module code. Possibly trying to port Flag module over in the process. Even getting that started would be an improvement. I want to own the 8.x branch of Flag. There's also some CMI issues I could poke at, as well as anything with a Novice tag that I feel I could do.