Mar 2006
Herzlichen Glückwunsch, Stefan!
I was happy to see that my other favourite blog describing a T2000 Test ended up a keeper - Stefan Rubner is listed as the newest Sun Fire CoolThreads Performance Contest Winner on the Try and Buy Programm.
I have no idea though, why they are linking to a single entry covering day 9 of his testing experience instead of the 11 Sunny Days that he’s been describing in detail up to now. So everyone who has only read that entry - don’t forget to catch up with the rest of his story.
I’m glad he gets to keep his testing equipment, and am eager to read more as he hopefully continues to explore OpenSolaris and the details of zones, zfs, … with that incredible toy now his own.
(That’s of course a perfectly egoistic desire - liking what I read on his testing, I had nudged our publisher to send him a copy, hoping to get some feedback along the way as he proceeds. ;-)
Now that a more reports from try and buy customers are surfacing, an incredible advantage inherent to the program is showing, that was somewhat hidden before. The program is not “just” about loaning an incredible piece of hardware that has the potential of reorganizing datacenters, with it comes a wave of support from inside and outside Sun.
As people are publishing their first tentative and often untuned results, their postings get noticed by a crowd of others interested in their findings, and from inside and outside Sun, advice, help, and genuine interest streams back to them.
In the end, everyone has won. The initial poster gains knowledge and advice on tuning his system, others see the testing expanded to cover their points of interest, and Sun got something from everyone - improved test results, “Niagara in the News”, and a living demonstration that their Systems are carried by a broad and supportive community.
There’s a new set of rules for the T2000 test drive “contest”. New at least that I didn’t notice them when I applied for the 60 day trial.
Update: Looks like others also noted this as new. See The case of the missing servers for a breakdown.
To me, the new rules don’t necessarily imply that there will be less systems given away, though, but I tend to like the previous wording better - while the updated version gives more clarity as to numbers, and might even result in more systems being awarded in total, the old version made the whole thing feel more like an award to be achieved.
From first reading, it sounds to me as if under the “new rules” no two systems can be awarded in the same month, even if there are two outstanding articles. It would seem a pity to me if one month a truly excellent article finishes second just to be followed by another month seeing a less convincing piece featured just for lack (or luck) of adaequate competition.
For the record, my own favourites for keepers are:
I hope the team reviewing the publications likes them as much as I do.
Update #2: Glad to hear that heanet got to keep the toy they’ve put to such a convincing performance.
Thanks to a pointer at Jörgs Blog c0t0d0s0, I stumbled across a very special Blog describing a T2000 test ride in detail. Special, since the author shares with us his experience from the point of view of an experienced Linux admin, describing his first steps into Solaris universe.
We who are so deeply familiar with the “Solaris way” of doing things often have developed a blind spot to other ways, and things that seem natural to us (and often even escape mentioning in documentation) may be a neverending cause of confusion to others. What a chance to learn!
As a side note - there’s another (not T2000 related) Blog I value for the quality of showing the “missing” links in documentation and explanation: The “unknown product manager” blogging there has made me reconsider quite a few topics.
From the comments in Stefan Rubners blog, it looks like the detailled “howto” descriptions recounting his progress along the way allow others to follow his steps and advance in their own trials. This is something other blogs, as valuable as they are in their technical detail, fail to offer. Excellent!
… and, a valuable marketing bonus for Sun: He’s reaching an audience Sun themselves have a hard time to reach - seasoned Linux admins and users who, if not for this detailled documentation, might not waste a thought on something as foreign as “another” OS.
Tomorrow it will be two weeks since I’ve applied for the trial, but so far there has been no reply or contact related to that application at all.
Right after CeBit I stormed the T2000 “Try IT Free for 60 Days” offer. It looked like a simple click-through-and-be-happy, but reading the License agreement I found that I couldn’t sign it in good conscience, which probably wipes me from the list of eligible recipients.
Why? I’m in rather obvious violation with the following passage:
4.3.Company agrees not to incorporate any Sun Trademarks into Company’s
trademarks, service marks, company names, Internet addresses, domain
names, or any other similar designations.
Apart from me not being a company, both the title of the book and this webpage use the name OpenSolaris, which is a Trademark owned by Sun. Especially the limitation “Internet addresses” most likely would apply to many others as well, who might be using Terms like Sun, Solaris, OpenSolaris etc as part of an URL.
All hope isn’t lost yet, the acknowledging page read “your responses to our qualification questions require us to review it further”, promising that “A Sun sales representative will contact you further”. So far, this hasn’t happened though, and I keep fearing that it won’t.
At the From chaos comes order blog I found a number of extremely useful tools when it comes to data center planning described, among them Power Calculators that are supposed to show a realistic figure for power consumption by the new systems.
I’m including direct links to these Power calculators here, but don’t miss the entry at that blog itself.
They’re configurable into nit-picking detail, did anyone compare the calculated values with measured ones to get a sizable error dimension?
While the around 70 Watts per CPU is something to get excited about, there’s no reason to hide the values for a fully stuffed system.
I’m afraid these systems won’t perform well as auxiliary heating.
On CeBIT, I found myself sneaking around the T2000 on display like a cat around a bowl of cream. Jörg Möllenkamp (who has blogged his CeBIT experience (mostly in German)) pointed me to the offer to try the beast for 60 days. Guess what was the first thing I did back home…