chuck.goolsbee.org

Just what has been going on at work the past week or so…

It is an MGE EPS 7000 300/500kVA UPS.

…to move a 6800lb (~3000kg) UPS?

Thanks to the amazing Hilman Rollers, only four and a half.

This is our new UPS at digital.forest. It is an MGE EPS 7000, which is a very cool unit. As purchased it is a 300 kVA system, but as we grow we can scale it up to 500 kVA. The battery cabinets arrive tomorrow. You can read about the UPS arrival on my blog at work.

I wrote a lengthy bit about communications as a key to surviving an IT disaster, which in many ways was a written version of the session I delivered at the MacIT conference at Macworld Expo last month. I tackle the stereotype of geeks as poor communicators, and lay out a strategy for getting IT departments into the communication habit. The stunning revelation that lead me down this road is a conclusion I came to when discussing an outage with a “layperson”… that is a user of technology rather than a maintainer of it. To him awareness was more important than downtime. Downtime didn’t bother him so much, so long as he was kept informed of what was going on, why, and when things would be back up. Forewarning would be even better. His downtime came about during a datacenter migration. A light bulb went off over my head, as I had successfully pulled off more than one datacenter migration within the past few years. Did everything go perfectly? Of course not, but the difference was that I put a huge emphasis on communication with our customers way before, before, during, and after the moves. I’m not some IT genius by any stretch of the imagination, and I’m not the first to use this tool effectively. It just seems that most IT professionals forget this critical part of their management strategy.

Anyway, for the terminally curious, the series is linked below. My editor wisely split it into two parts.

Part One

Part Two

The server that hosts a lot of my images is down. It is my own personal box, which is almost as “vintage” as the cars it displays. It is a wonder that it works at all to be honest… clean living in a clean room I guess.

It had a minor disk issue earlier this morning and I’ve fixed it, but now I’m making a backup before I bring it back online. “Have patients!” …said the Mad Doctor!

Update: OK, as of 11:45 Pacific Standard Time the image server is back online.

OK, so maybe somebody at Apple has a clue or knows how to listen. They announced a new rev of the Xserve today. I won’t bother to talk about the stuff everyone focusses on (CPU horsepower and whatnot, I have friends and customers you can turn to in order to get the skinny on what’s happening inside the new box. ) I’ll stick to the subject of all my usual rantings about servers and server design, the case. This is because I don’t manage servers, as in “what goes on inside the server” I manage Datacenters, namely what happens OUTSIDE the server once it is racked and operating.

The momentous cause of my small celebration today? Apple put a USB port on the FRONT of the Xserve. Whoo hoo!

Mind you this is only a very small step away from ’style” and towards “substance”, and ironically “usability” but it IS progress and I have to give Apple credit for that.

As I have said before, to be truly useful in the environments it was designed for the Xserve should have all “user” ports on the front, namely USB, and Video, and all “system” ports on the back, namely power, network, FibreChannel, etc. If it connects to another system or the datacenter infrastructure, it goes on the back. If it interacts with a user, it goes on the front.

Datacenters are laid out in hot aisles and cold aisles, where the hot back sides of servers are isolated from the cold intake side. This allows for optimum cooling and airflow. In ideal datacenter environments the hot aisles will be contained and the heat given a specific path for removal. If users have to constantly have access to the back side of racks (or more accurately the hot aisles) then they can not be easily contained. Putting user-required ports on the back side of servers is counter-productive.

Of course, that isn’t my biggest complaint about the Xserve’s design. That remains the completely absurd overall length of the box, which still lays out to 30″ (76.2cm) which is so long that it completely obliterates and density advantage a 1U server supposedly buys you.

I know I’ll get video ports on the front panel long before Apple pulls their head out their butts on case length of 1U boxes though.

Thanks guys.

Note: This is something I zipped off just now to “Server Specs: A SearchDataCenter.com blog” after I had a conversation with Matt Stansberry via IM. He was bugging me about not writing anything for a while, and I claimed how busy I was. Then we were just casually conversing about some industry trends when I made my “ramps and staircases” analogy. He said something like: “If you had typed that in WP instead of IM it would be published!”

True enough. So Sunday before we went out to get our holiday tree I cranked this out. After sleeping on it I added a few bits, then had a couple of trusted friends (writers both) rip up my grammar and wording and put it back together looking much more polished (Thanks Bill & John!) So if market analysis is your thing click “more”…

(more…)

Heard from several folks over the past few days asking if we’re under water. We’re fine… high and dry… both at home and at the office. Ironically the office park where digital.forest lived from 1997 until 2005 DID get severely flooded, and at least one of the datacenters in that area (T-mobile) suffered an outage. I’m SO glad we left North Creek Parkway. I miss the paths, and I miss the salmon in the creek, and I miss Teriyaki Etc., and more than anything, I miss Hobin-san and Sato-san at Hana Sushi… BUT I’m so happy to be at this amazing facility in Seattle. (But I digress)

Despite what the news media says, these things happen every year in the Autumn. Usually in November, but it can happen anytime from Mid-October until Christmas. Low pressure sits off Vancouver Island, pushing the jet stream south, pullling cold air down from the north… we get snow all the way down to sea level, and the mountains usually get DUMPED on… several feet of snow in a matter of a day or two. Then just as suddenly, the low moves on and the jet stream snaps back north pulling warm, very moist air from the Pacific. The traditional term for it is a “Pineapple Express.” All the snow melts in a matter of hours, sometimes minutes. It takes the better part of a day for it to come down out of the mountains but if it continues to rain while that is happening the lowlands start seeing floods.

Last year it happened in early November.

This year the temperature change was swift and dramatic. Sunday night I was up late brewing a big batch of BioDiesel. I was acutely aware of the temperature (in the mid 30s F) because I was wearing my big winter coat and I also have a big temp dial on the processor. When I pumped my waste oil into the processor from the settling tank it moved like molasses. I looked at the temp gauge and it read about 34°F (or ~.5°C for my worldly Metric readers.) There was frost here and there around the barn as well, and the gravel of the driveway made that crunchy sound like ice when I walked up it to the house. I awoke and ventured out to the barn at 5 AM to shut off the processor… throwing my big winter coat on as I prepared to leave the front door. As I stepped off the front porch my brain registered something odd. Half way down the driveway I realized what it was… “damn, it is WARM out here.” There was a light rain, and some wind, but the temperature was way up from the night before. Close to 60°F/15.5°C. I shut off the processor (temp gauge there read ~100°F/38°C… about right for a batch of BioD that had processed for several hours) and walked back to th house with the coat in my hands. Checked the weather widget on my OSX dashboard and it reported 58°F/15°C… so my guess was right.

10-12 hours later the floods began down below… like clockwork.