[content-discuss] Re: Article Idea
David J. Orman
ormandj at corenode.com
Fri Mar 31 00:24:29 PST 2006
Jim,
I just wanted to follow up and say I agreed with all that you said.
I've been extremely busy so haven't had time to write an appropriate
response, but rest assured - I will. I think your idea of multiple
articles makes perfect sense, and I am inclined to take this route.
I'm glad you like my method of writing, and I will be more than happy
to dedicate time into providing you with GOOD write-ups. I'd rate
that short blurb of mine as a 2 or 3 at most, it was my fingers
flying as my mind ran rampant, there was no thought given to actual
delivery and composition. ;) Actually, there might be something said
for that. I'd like to combine the two. Proper English + spur of the
moment thought. As to the interview, I'd be more than happy to take
that on. Your completely correct in saying it will bring added value
to things. My perspective + an outside perspective == closer to the
truth. That's what I hope your goal is, and I know what my goal is.
Providing everyone with information that truly benefits them and
helps them in their day to day operations.
I will elaborate more as I have time - it's 2223 here now and I wake
in 4 hours and 37 minutes, so bedtime for me!
Cheers,
David
On Mar 27, 2006, at 9:16 AM, Jim Grisanzio wrote:
> David J. Orman wrote:
>> I'm here as well, so please feel free to communicate with me via
>> this list so we can keep everything documented and get multiple
>> opinions/ feedback/input/etc. :)
>
> Welcome! We formed this project a couple of months ago to generate
> some content for the project -- articles, how-to's, personal
> accounts, interviews, Q&As, case studies, profiles, etc. Text is
> great, but I'd like to eventually grow into having podcasts, screen
> casts, and video as well. This content we produce will live all
> over the site, of course. Some stuff will live on the articles
> page, some within other communities, and some off site altogether.
> Personally, I'm most interested in the people that produce this
> technology, but others may be more interested in the technology
> itself. We need both, in my opinion.
>
> The efforts here are 100% community generated. Sun is not running
> this Content Project, in other words. Our resources are only what
> we as individuals bring to it. Our process for publication are
> designed to be as lightweight as possible, with the key being peer
> review. If an article is highly technical, it needs to be reviewed
> by the community before publication. That's the value of publishing
> on opensolaris.org rather than a blog or somewhere else (though
> those venues certainly have value but different value).
>
>> PS - I need to make clear now before a lot of time gets invested
>> that we are still in the evaluation phase of a _lot_ of things.
>> We are truly a "fledgling" data center as described on our site.
>> There are a lot of things still being ironed out. My facility is
>> still far from being finished!
>
> Cool. I think many people are in your position and by you
> documenting your experiences for us will help other community members.
>
>> We're just trying to start with the right technology from the
>> beginning, and design our environment around the "stuff" that
>> makes everything go. Not try to shoehorn the "stuff" into a
>> poorly designed environment. We've got a 12,000 sq foot
>> underground facility, all concrete walls. We're building it out
>> by ourselves, and with funding from our day-jobs. We don't have
>> millions of dollars, we don't have paid construction workers.
>> Everything from sandblasting to demolition to engineering
>> (electrical, structural, you name it) we are doing in-house.
>
> *That* is why this story is interesting.
>
>> I'm a 24 year old college graduate from Dallas,
>
> Now I feel sick. :) I *used* to be 24. :)
>
>> I moved to Hawaii with the goal of starting a data center,
>
> Back two or so years ago when we were planning OpenSolaris, we
> listed many reasons why we wanted to build a community. Engaging
> entrepreneurs was very high on the list.
>
>> as on a trip here prior to my graduation, I saw there really
>> wasn't much in that sector in this area. Our only decent sized
>> competitor is full, and turning away customers. The other two
>> competitors are priced insanely high. Bandwidth is expensive in
>> Hawaii, but not _that_ expensive.
>> Most of the companies around here have "old" technology. Most do
>> not have backup locations. The pre-existing data center owned by
>> Equinix here is at/below sea level, right next to the airport.
>> Etc etc. Just giving some background on the logic behind my
>> coming out here.
>> We've been in touch with a lot of potential customers (the City
>> and County of Honolulu is coming out in April) .. I've got a
>> very interesting auditor report on their main data center. You'll
>> love the indoor rain gutters and "paper towel water detection
>> system" they use in the walls.
>> Needless to say, we've gotten a LOT of interest in our services
>> since our publication in Pacific Business Journal. Now we're
>> working on figuring out what we want to base all of our services
>> on. I will go more into my exposure to Solaris/Sun later, but
>> needless to say I ended up exploring Sun as an option. It didn't
>> take long to very quickly realize I had something good in my
>> hands. We've evaluated Linux (various distros), FreeBSD, and now
>> Solaris (still in evaluation.. actually.. I'm just playing and
>> having fun like a little kid!) Linux was turned down due to
>> pricey support and generally ad- hoc thrown-together design (from
>> my standpoint), as well as performance issues. It "felt" like
>> using a duct-tape solution, not something industrial grade. That
>> might be fine when you're running a cluster and have designed
>> failover, but that really isn't that efficient in power/heat
>> production. As I'm sure most of you are aware, that is the
>> largest cost a data center has, electricity. PS - I am keeping
>> this very short/abridged, more details can be provided as needed.
>> FreeBSD was fairly nice. This was to be our final solution, until
>> I decided to try out Solaris. I never really had any issues with
>> it, it seemed more "put together" and I felt a lot more secure
>> using it as our primary OS. We were running a solution for our
>> needs with apache/ mysql/pgsql/postfix/dovecot/proftp. It was a
>> pretty nasty little combination, but it worked well enough for
>> our needs. (As a side note, we are about to start evaluating JES.
>> My initial experience with it was terrible too, but I'm told once
>> you wrap your head around it, it's really nice.)
>> I then had a client come and ask me (in the past few weeks) for a
>> backup solution price quote. I said sure, his IT guy got in
>> contact with me and provided me requirements. 25 gigs initial
>> storage, and growth of about 5-10 gigs a month. The thing is,
>> they wanted daily backups for a week, weekly backups for a month,
>> monthly backups for a year..
>> I got a slight headache at this point, as the client made it
>> quite clear he was looking for an inexpensive but reliable
>> solution. I was then informed he had a business DSL connection.
>> Another big uh-oh. _Then_ I was informed (here's the bomb) they
>> wanted full backups, because they wanted to easily be able to
>> retrieve any day/week/month/ etc worth of files at any time,
>> without *our* (my company's) intervention. I was nearly in tears
>> trying to think of a solution that could do all of these things,
>> and not cost the client thousands a month.
>> That is when it dawned on me. That cool Sun feature slated for
>> Solaris 10! What's it called? Oh yeah, ZFS. I think I remember
>> something called snapshots, let me give it a go!
>> What followed was about two weeks of tears trying to build a
>> machine that would run Solaris (and not drop from the network
>> every ten minutes.) As I said, we're funding every aspect of a
>> data center from our pockets, blood sweat and tears, working day
>> jobs. Needless to say, buying Sun gear was not an option for us
>> without the assurance of a customer's pay coming in. We also did
>> not want to take the "try a coolthreads server free for 60 days"
>> deal, because we felt it wouldn't be fair to Sun for us to take a
>> trial if we had absolutely no idea if it would even work for us.
>> At the same time, we needed something to try out this fancy new
>> ZFS filesystem. After managing to source parts that would work
>> with Solaris (no easy feat here in Hawaii!) and getting b33 of
>> Nevada installed, I was blown away.
>> I needed help getting things working properly, but *without a
>> support contract* and running on *completely unsupported off the
>> shelf hardware* - to my amazement, Sun engineers responded to
>> simple and basic (ignorant) mailing list posts I made. I got
>> clear concise answers to my problems, and quite quickly! Within a
>> day, I had the client's solution already proven technically
>> feasible, and not only that, but INEXPENSIVELY and RELIABLY. Want
>> backups from 03/15/06? No problem, just browse to .zfs/snapshot/
>> 031506/ and you're there. Grab whatever you need. Have fun!
>> Us? Not wasting a ton of storage on full backups of that amount
>> of data. Sleeping soundly at night not worrying about the data
>> being safe. Very little equipment to manage/run the system, so
>> low power costs. Capability to grow to as much storage as needed.
>> Best of all, very little time invested! Literally, from question
>> to solution, removing all time required to evaluate other
>> solutions and source parts, was no more than 6 hours. This coming
>> from a guy who never used Solaris before the past few months. ZFS
>> is so intuitive, it was like a breath of fresh air. Sun engineers
>> filled my mailbox with tips, hints, suggestions, explanations.
>> Even suggestions on drive cages!
>> Who do you think my first server purchase will be from? You got it
>> - Sun. As soon as things are ironed out with this customer, we'll
>> be picking up a nice Sun server to top that drive cage and export
>> the ZFS share via NFS. Possibly another Sun server to handle the
>> frontend operations. It all depends on budget, of course! This is
>> just the start, just the core, of what we plan for the future.
>> It's just a few days of my life, and my eye-opening to Sun.
>> Don't even get me started on Niagra. We're looking at that now
>> too, and realizing with the savings we'll have in power and
>> space, we'll be competitive in price with data centers were never
>> thought we could be, due to high bandwidth pricing in Hawaii.
>> Not only that, but we'll be able to offer solutions nobody else
>> can here, due to all of the innovative technologies in Solaris
>> 10. We're starting clean, we can implement everything around the
>> solution that best fits us. From the looks of things, that
>> solution is Sun, and we're building our data center around that
>> core. :)
>> Cheers,
>> David
>> PS - This was just to give some background, I didn't write it
>> with the intention of it being used directly. I can write up
>> something more formal and concise as needed. I just thought it'd
>> be nice to share a little more into where we stand, where we are
>> going, and why Sun looks like our best option to get there. I'll
>> let you all take the ball and roll it around for a bit. Feel free
>> to shoot back any questions, suggestions, feedback - anything.
>> I'm here to provide anything I can. :)
>
>
>
> Damn, this is excellent. I'm not a technologist so I have zero
> feedback on the specifics of what you are putting together. Over
> time, people on this list and on the discuss and other lists will
> offer more technical feedback. I'm interested in the initiative
> itself, in how you did it, why you did it, the successes and
> failures, the challenges, the financing -- the human aspects of the
> business, basically.
>
> So, we need to figure out a mechanism of producing this document or
> series of documents. I like your writing style, so I think you need
> to write some or most of this, but I'd also like you to consider
> someone interviewing you as well. The reason is this: the
> information generated when people write is many times different
> (yet complementary) to the information generated when they are
> interviewed.
>
> Since you are in the evaluation phase of your operation, you have
> time to map out a series of articles. How about this for a
> suggestion: a general overview article to set the stage for the
> case study, then two or three articles to dig down into the
> technology itself. Maybe one is on the technology, one is one the
> business issues, customer issues, etc. In these supplemental
> pieces, you dig down very deeply. Then, to support all that,
> someone interviews you and does a Q&A. You just relax and answer
> questions and let things flow wherever they go.
>
> You can write this stuff and post drafts on the Content Project. An
> outline should probably be scratched out first, but please feel to
> be as informal as you like. When you are ready for a formal review,
> we'll post the draft(s) and call for review on the specific
> technologies you mention. Then you update your draft as needed and
> the reviewers are credited as contributors as well. The more
> detailed the article is, the more review it should have so it hold
> ups up over time. Your experience is yours, of course, but the
> articulation of specific technologies being reviewed would be
> valuable to the community. Now, the Q&A would require a lot less
> review, of course. We'll need to consider each one individually.
>
> If you choose to do something like this, we can roll it out over
> time. Or if you are more comfortable putting it all in one article,
> that's fine too. I'd love to hear your opinion and the opinions of
> others. Once the content is generated, however, we can all blog
> about it and draw attention to it and I'm sure more conversations
> will start as a result.
>
> Just some ideas .....
>
>
>
>
>
>> On Mar 23, 2006, at 10:28 PM, Jim Grisanzio wrote:
>>> I'm forwarding part of this thread:
>>> http://www.opensolaris.org/jive/thread.jspa?threadID=6964&tstart=0
>>>
>>> Bill suggested that David's situation may make a good article.
>>> David seems to agree. I do, too. So, I thought you'd like to see
>>> some of David's thoughts and my ideas for a piece.
>>>
>>> Jim
>>>
>>>
>>>
>>> -------- Original Message --------
>>> Subject: Re: [nfs-discuss] New to ZFS/NFS - Question concerning
>>> permissions
>>> Date: Thu, 23 Mar 2006 18:25:15 -0800
>>> From: Jim Grisanzio <jim.grisanzio at sun.com>
>>> To: David J. Orman <ormandj at corenode.com>
>>> CC: bill at rushmores.net, nfs-discuss at opensolaris.org
>>>
>>> David ... cool. Let's do it. I'll forward your mail to the content
>>> project alias ( content-discuss at opensolaris.org ) and we can
>>> talk about
>>> processes, review, etc. Anyone interested in participating,
>>> please feel
>>> free to jump in. I think a piece like this would take some time, it
>>> would probably need some artwork, and it would certainly need
>>> technical
>>> review.
>>>
>>> Jim
>>>
>>>
>>> David J. Orman wrote:
>>>
>>>> Jim,
>>>> It sounds like a wonderful idea, maybe it would also help other
>>>> people see
>>>> the advantages that using Sun systems could bring them. I have
>>>> a lot of
>>>> information I can share concerning my movement forward in this
>>>> area, from
>>>> the first mention of Solaris to me, until the present.
>>>> Everything from
>>>> failures to successes, and plenty of surprises too.
>>>> I'd be happy to help out however I could, with the amount of free
>>>> "support" I've gotten from the OSOL community for my
>>>> *commercial* venture,
>>>> I feel a sense of responsibility to give back at least as much
>>>> as I have
>>>> recieved!
>>>> That is the beauty of an open development model where the
>>>> engineers who
>>>> are normally delegated to cubicals with no social interaction
>>>> are let lose
>>>> with the people who use the software/hardware they write. :)
>>>> There is a
>>>> symbiotic relationship between the producers and the consumers,
>>>> and they
>>>> both influence each other in a positive fashion. The producers
>>>> learn what
>>>> the consumers *really* want, and the consumers learn how to use
>>>> what the
>>>> producers created to the full extent. No longer are you reliant
>>>> on surveys
>>>> and other flawed statistical analysis to get an accurate
>>>> picture of your
>>>> customer base.
>>>> Solaris has a lot going for it now, and especially when S10U2
>>>> hits the
>>>> shelves it's going to be a solution that solves a LOT of
>>>> headaches. With
>>>> the SarbOx requirements coming due soon enough, storage needs
>>>> will be at
>>>> an all-time high. Data integrity and validity will as well. ZFS
>>>> makes a
>>>> perfect solution for this, and you can have it for *free*. Maybe
>>>> a simple
>>>> case study will be enough to show companies how this can impact
>>>> their
>>>> business in a positive fashion. Just a few thousand dollars
>>>> worth of Sun
>>>> equipment later and they to can be on their way to sleeping
>>>> peacefully at
>>>> night when the SarbOx enforcers come knocking on people's doors. I
>>>> wouldn't want to be the tech guy at a company who doesn't have
>>>> something
>>>> like ZFS handling their data store.
>>>> I'm a long-time FreeBSD user, and prior to that a Linux guy. I
>>>> came upon
>>>> Solaris by accident, and after my initial trials I was very
>>>> unimpressed. A
>>>> lot of that has changed in the recent past. Like most things,
>>>> it takes
>>>> some effort to see the advantages when you don't really even
>>>> know what to
>>>> look for. Maybe a nice writeup of real-world planning, analysis,
>>>> and
>>>> implementation could help people save a lot of time. :)
>>>> Anyways, I think we are deviating from the original topic, so
>>>> if you'd
>>>> like to continue this further, direct me to the list or contact me
>>>> directly via this email address.
>>>> Cheers,
>>>> David
>>>> PS - Cool, I've never been in a blog before! I feel special!
>>>
>>>
>>>
>>>
>>>
>>>
>>>>> Hello, David.
>>>>>
>>>>> Bill Rushmore pointed this out to me in the Content Project
>>>>> today, and I
>>>>> think it could make an interesting Case Study -- need, evaluation,
>>>>> implemention, Sun systems, Solaris and OpenSolaris community
>>>>> elements,
>>>>> the whole thing. Would you be interested in considering that
>>>>> for an
>>>>> article?
>>>>>
>>>>> We have a Content Project going here:
>>>>> http://www.opensolaris.org/os/project/content/
>>>>>
>>>>> And we are starting to post peer-reviewed articles here:
>>>>> http://www.opensolaris.org/os/articles/
>>>>>
>>>>> Perhaps this could be a Case Study that you write. Perhaps it's
>>>>> something that someone else writes by interviewing you and
>>>>> digs into the
>>>>> technology issues, etc. I see Peter Harvey blogged about it,
>>>>> so maybe
>>>>> he'd be interested in writing a longer article about the
>>>>> system you end
>>>>> up building and how you built it.
>>>>>
>>>>> We can talk about the specifics of the piece later, but I just
>>>>> wanted to
>>>>> float the idea to you guys to see if there was some interest.
>>>>>
>>>>> Jim
>>>>> --
>>>>> Jim Grisanzio, Community Manager, OpenSolaris
>>>>> http://blogs.sun.com/jimgris/
>>>
>>>
>>>
>>>
>>>
>> _______________________________________________
>> content-discuss mailing list
>> content-discuss at opensolaris.org
>> http://mail.opensolaris.org/mailman/listinfo/content-discuss
More information about the content-discuss
mailing list