150TB write life. What. How is that remotely adequate?
I have all of the sads. Every single sad.
6309 posts • joined 31 May 2010
150TB write life. What. How is that remotely adequate?
I have all of the sads. Every single sad.
Right, so even you know it's not true and that isn't my position.
No, I think you're capable of compartmentalized thinking. You seem perfectly capable of believing one thing but rabidly espousing another. Typical of religious types, actually.
All I was hoping for was that folks would spend 5 minutes having a look at plan B because it may work out well for them. I don't think that counts as zealotry and I don't really think it's worth having a flamewar over either because it's common sense.
Which you failed to make a case for, especially given the timeframes involved, all while freaking out because extreme edge case scenarios weren't given equal billing to realistic solutions for the overwhelming majority.
It is you who is behaving like a "religious wanker" (insults, misrepresentation, pretending you know what other people think, asserting you know best with zero evidence to back it up, intolerance), and as a rule dogmatic loudmouths don't tolerate competition, so that comes as no surprise.
I love competition. It keeps us all on our toes. If and when any shoes up, I'll gladly engage with it. I've proven that by cheerfully having comment thread discussions that have gone on for dozens of comments about different ways to solve problems. Where the commenters involved are looking to actually debate (as opposed to proselytize) I genuinely enjoy such discussions. They're the reason I read El Reg.
As for "knowing what you think", I haven't a clue what's going on in your mind. I do know what you are saying in your comments. And what you are saying in your comments is completely disconnected from what you claim to be thinking, so either you are absolutely awful at expressing yourself or you are engaging in some pretty perverse compartmentalized thinking.
... because being around someone who misrepresents folks and then flames them on the basis of that misrepresentation isn't fun, it's just plain old bullying and bullshit.
I haven't misrepresented what you are saying at all. In fact, I'm decreasingly sure that you're even sure what you are saying. You want people to think about Linux as an alternative, that's fine. Nobody here will give you shit for that. Everyone in this entire comment thread agrees that Linux is a viable alternative that should be considered when and where possible.
But you don't end there. You have gone off into crazy freak-out land about the fact that the article didn't present Linux as a viable enough alternative which is completely false: I gave Linux far more column inches than it is due, given the insignificant number of workloads that can be reliably migrated in teh one month timeframe under discussion in the article.
You seem to believe that I - and others - are somehow bashing Linux, saying Linux shouldn't be considered and otherwise pissing in your religion-flavoured cheerios when absolutely fucking nobody here is doing that. The discussion we're having is about when and where it's viable, in what timeframes and for what workloads.
You're acting as though we killed your dog because we aren't championing Linux as the first, last, and only solution regardless of the viability, practicability or you'll-get-fucking-sued-for-taking-those-risks involved. Which is nuts. Wonko. Loony Tunes.
Your whole approach to commenting here has been off. You have basically just walked up and taken a dump on my lawn, claiming you did so because you're just standing up for the beleaguered natural fertilizer industry, whom I neglected to adequately endorse in my article about how to winterize your perennials before the first frost. Completely disregarding the part where two paragraphs of the article were devoted to the fact that for some plants it's a good idea mix in some natural fertilizer for the overwinter process.
Telling you to stop shitting on my lawn isn't bullying. Chasing you off my lawn with a rider mower would be.
He's not giving you his condescending tone, get past it and move on.
For one, there's be a lot more creative use of invectives, were I actually using my naughty voice. ;)
Thanks for the clarifications, I think we just have different SME client experiences
Well, this is why in my original comment I said that I, personally, was reaping the benefits of getting away with bypassing some change management procedures and put a smiley on the end. Just because that setup works for me doesn't mean it works for everyone. I take the time to learn a lot about the business, and have the fortune of being supported by some really plugged in tech weenies on the local sites.
Also to me "Taking the financials server down for an hour at 3am to hump it up a few Windows versions is not a major outage." would be a major outage with one of my client's, as their Financials is a full MRP system and hence is only supported on a restricted suite of platform software.
Well, yes. Each application is different. Which is sort of my point in the whole conversation in both articles and both comment threads. It's also sort of endemic to the discussion about the ease with which a sysadmin can (or can't) bypass change management.
I know my workloads. Inside and out. I've worked with them for over a decade in most cases. So there are lots of them that I can get away with doing "unscheduled" maintenance if I keep that maintenance to the right outage window.
By the same token, there are a bunch where if it's down for more than a minute or two, I will be shot in the streets.
That's the key here though: the diversity of IT workloads. It's why there is no one response that fits. It's why there is no one set of procedures that efficiently encompasses all endeavors and why migrating to Linux in a month is absolutely impossible for the overwhelming majority of workloads. Hell, migrating from Windows to Windows in a month will be a stretch for many workloads.
Despite that, there are a significant number of workloads where simply humping it up a Windows version is no big deal. One of the measures of a sysadmin is to be able to know into which category their individual workloads will fall.
But the main point which I think we agree on, is that you need to have some form of change management process which gives you a framework in which to make such decisions and communicate appropriately, even if at times the only business visible output is a line item on an invoice.
I don't disagree in general, but I think that all change management processes need to be fairly flexible. As workload and maintenance types vary we need the ability to bypass the more bureaucratic and restrictive bits when and where it's sane.
Where I get my panties in a bunch about change management is where organizations go full bureaucrat.
This one time, while consulting (but not at band camp) I was summoned to solve a problem with an application on an end user device. The device was managed by an outsourcing company that was itself managed by another outsourcing company and five layers up with the consulate of a very large and powerful nation. The endpoints were ancient. The printers were ancient. Everything was ancient and everything was locked right the hell down.
I looked at what was asked, looked at the permissions that had been given my user account to accomplish them and realized that it was impossible. So I booted off of my USB key, loaded the registry hive on the endpoint, made the changed by hand and it worked. The consulate staff got on with their day.
I then spend 16 consecutive hours on the phone with different layers of support and management to get official permission to have my user account granted the rights to do what I'd already done so that it can be signed off on and properly accounted for.
I probably violated about a dozen treaties and maybe some laws in doing what I did. I also saved the day. That country's ambassador actually showed up in the consulate while I was busy playing ping-pong for authorization with support and ended up having to make emergency use of the computers to use the very application I had enabled.
I found out later that the result of that conversation was that the country in question extracted some people from a city that was hit by a tsunami and brought them here to Canada. The Ambassador was relaying the results of his negotiations for refugee status for these folks and his aides sent over all the info on where these people were to live.
The reason I had to get the application working was because the new rules said that certain types of official communication couldn't occur without the new communications software installed. The Ambassador (who normally isn't in our city, so it usually doesn't matter if that software works or not) would have been restricted to firing off information via some form of certified snail mail otherwise. That could have slowed the country in question from moving those people, and who knows what that would have meant for them?
It was that incident that made me appreciate the need for flexibility in change management processes. The more rigid and absolute they become the less the serve the needs of the organization implementing them.
Change management needs to be change managed. :)
There will be plenty of people and companies for whom migrating from WS2003 to Linux (or something else, if you wish) is a perfectly feasible choice. But I don't see that given any real credence in the article. Hence it is rather hard to take seriously his assertions that he is unbiased when it comes to Microsoft stuff.
Not in a month, there's not. I gave exactly as much credence as is rational to the idea of moving to Linux in a month in both articles. There are absolutely some workloads that can be moved in that timeframe. I even discussed some candidates. But the overwhelming majority of workloads cannot be moved in a month.
You just don't seem to be able to understand that. The very specific topic of both articles is the month remaining. The possibility of moving some workloads to Linux is discussed. The reality that most can't be moved in the timeframe given is also discussed.
Given a longer timeframe - and enough money - anything can be moved to Linux. I don't anywhere attempt to say that's a bad thing, or an impossible thing. But I am not going to pretend like Linux is a viable option for most companies facing that ticking clock when it isn't.
As I said in the article: do the right thing and move your Windows workloads over to a supported version of Windows. Then, once the immediate crisis is passed you can look at moving those workloads to Linux, where there is time (and hopefully resources) to do the job properly.
And if you want to get into a pissing contest about who has worked on the smallest networks or those with the least resources, well...
...I'll fnording wreck you, punk. Cash strapped SMBs with are my sandbox.
I've spent my professional career moving every workload I can get my hands on over to Linux, Solaris and BSD, when and where possible. I've moved everything from overloaded, overworked everything-on-one-box systems to entire render farms. I'm probably one of only a handful of people seriously doing Linux VDI.
I have a setup in a northern Albertan village where 9 companies are running their IT infrastructure off a single Netgear WNDR3700V2 router with a RAID array attached to the USB port, because it's all they could collectively afford, but it runs external facing websites that customers around the world use to purchase from these companies as well internal order and process tracking systems. It handles backups, networking and everything else except the dumb end-user terminals, which are recycled Wyse clients that do nothing but run Chrome on Windows XP Embedded.
I know about migrating to Linux. Big and small. And you know what? The overwhelming majority of workloads on Windows Server 2003 systems can not be reliably migrated to Linux with a month left on the clock. Period. There isn't room for debate on that.
That you somehow need to take that very narrow statement - which is what both articles were about - and somehow expand that to encompass more is your problem. If you want to change the timeframe, then the discussion changes. Linux migrations become more feasible the wider you make the migration window, and the more money you make available to do the migration.
I qualified my statements just fine in that article. Maybe you'd do well to actually read it.
What I am saying is that more and more companies have switched to Linux and now they are very much enjoying the benefits.
And? What has this got to do with the price of rice?
It appears that Microsoft are very good at locking people in with their propitiatory systems
Who here is debating that?
and for this reason, it is not always an easy task to switch, but you can escape them if you try
Who here is arguing against that?
(not sure why we need to impose an arbitrary time limit of 1 month though?)
Because Windows Server 2003 support ends in a month. This the time limit that is endemic to the very narrowly focused topic of two separate articles in which the extremes of the Linux zealots can't seem to put aside the evangelism long enough to comprehend.
The articles were not about "don't migrate to Linux, ever". They were about "there's 1 month left on the clock, here are your realistic options".
I'm all for moving to Linux where possible. With one month left, it's only possible for an infinitesimal fraction of workloads.
I see no such demands in my post
Of course not. Fox News proclaims they are merely providing "balanced reporting", too. They aren't.
All I am trying to say is I think your position is too absolute, too black or white.
Because it is black and white. There is the overwhelming majority of use cases - > 99.9% - where everything I said in that article applies without question, and there is the extreme edge cases - < 0.1% - where it is advisable (or hell, possible!) to migrate workloads from a Windows environment to a Linux one in a month without issue.
If this is a grey topic it is so very slightly grey that even a remarkably precise, highly calibrated colourimiter would call it "white".
By return you have reiterated your point that you think that's the only valid approach
I have stated why I think it's the only valid approach for the overwhelming majority of cases and invited you - or anyone else - to provide concrete evidence to the contrary. Not for extreme edge cases, but for enough use cases to make a noticeable dent. Say, 10% of use cases. Bearing in mind the one month time frame.
hurled some insults and declared anyone to differ with your opinion to be incompetent
No. I called you incompetent. Because you differed with my opinion but couldn't back up your quite frankly religious sounding zealotry. You make a bunch of wild claims about how people should be able to move to Linux, but don't discuss how this will be accomplished in the time frame given, nor how to solve the very real issues that real world systems administrators will face.
And yes, I'll hurl insults over that. I have no time for religious wankers of any description.
You aren't sounding like much fun to work with.
I'm not. IT isn't about having fun. It's about getting a job done. A difficult, complex job.
I do the job and then I get paid. Go run your little world.
In small business'es the nerd's if there are any and the on-site IT support person have generally little understanding of the business
You'd be spectacularly wrong about almost every single SMB and midmarket organization I've interacted with. The nerds are integral. They help set business strategy and due to the tight integration and automation have to have a seat at every major discussion that goes on, because they'll be the ones implementing it. Next?
So asking the business, means that I don't shutdown the RDS server on the morning when they are using it to complete a major bid application before it's deadline...
And now you're way out in left field making completely unwarranted, wild assumptions. I pretty specifically stated in one of my previous comments that I tend to do notice-free upgrades late a night. Think 3:00 am, so that it's after backups run. I know the businesses I work for, and I choose times to do patch runs, server outages etc so that the business doesn't even notice that something is occurring. And yes, I check IM to see if anyone's on, even if it's 3am.
Major outages need to be scheduled in advance. Taking the financials server down for an hour at 3am to hump it up a few Windows versions is not a major outage. And, I should point out, I'd probably have tested it by humping the reports server up a few days prior. Again, without feeling a need to let people know.
Why? Because I can do the move and do the testing all before it affects anyone or anything and have time to revert if it all goes sideways.
In fact, it's usually the scheduled stuff that goes horribly wrong, because it's complicated enough to need to schedule it and there's never quite enough time to get something that big done before people start coming in for the morning.
I'm not sure what you mean by "Proper change notice"
Well, actually you nailed it. Things like Prince2 change forms are a great example of "proper" change notice. Long, drawn out bureaucratic cover-your-ass and make-it-look-like-you're-important exercises. They have their uses, but they emphatically don't need to be used for every minor change. Especially now that we are entering a world of dynamic, self-healing datacenters; $deity help you if you ever start using software defined networking!
In 95% of cases there is zero point in attempting to explain why I am "approaching change in a specific way". The business side wouldn't understand it and - to be more blunt about it - they don't care. I am not going to fill out a Prince2 form to renew an SSL certificate or push out security patches. Not going to happen.
The other 5% of changes are the ones that actually affect end users and customers. Here documenting and justifying everything is critical.
Server upgrades can fall into either category. I am not issuing a change request if I start rotating nodes out of an ESXi cluster for patching, so long as the cluster supports N+2 node failures. That means I can patch and not affect the ability of the cluster to lose a node during production.
Similarly, in the case I documented in a previous comment wherein a Win32 application can be migrated smoothly and easily between Windows versions I see zero reason to let anyone other than the local nerd know. The end users aren't affected at all; they don't interact with the operating system at all. They interact with a Win32 client that if very binary about things. It either works (it can access the database and the file share with all relevant permissions) or it doesn't. I'll know that after a few seconds of testing.
Change management can add value, but it should not be done simply for it's own sake. Each organization needs to examine exactly what level of detail they feel is required to meet the business' appetite for risk versus its ability to conduct operations in a timely manner.
Just as you've seen contractors who disregard any notion of sane change management (taking down RDS at 11am would be a great example), I've worked with many a company whose change management processes were so smothering and opaque that they prevented forward motion of any kind.
Small businesses rarely have the resources to go full bureaucrat on change management. The increase in staff costs of going full bureaucrat on anything is mind boggling.
You are coming across all "Trevor's Way or Highway"
I'm sorry that you have a problem with your brain where you perceive reality to be "Trevor's way". That's really weird.
"when in actual fact the world doesn't revolve around Windows"
I believe the earth rotates around a solid nickle-iron core doped heavily with radioactives that sunk during the late heavy bombardment.
I'm quite happy for you to state you think there are no options, but I have seen cases where there is no option but to move off Windows
No, what you want is for me to present marginal edge cases as being just as common as the overwhelming majority of cases, or of equal import/likelihood. That's bullshit.
I refuse to be Fox News. Fox will do ridiculous things like present climate scientists side by side with a climate skeptic and pretend that their views are equally valid and equally supported by evidence. They manufacture controversy by completely disproportionately weighting discussion and doing it over and over and over and over and over.
This sort of crap is known to work. It is part of the whole concept behind activities like affirmative action: make minorities disproportionately represented and they will seem like "the new normal", hopefully closing down one avenue for the spread of prejudice.
I flatly refuse your demands to undertake affirmative action regarding open source and present it's relevance disproportionately to reality.
We can argue about the scale of the last category 'til the cows come home, but the point is that category does exist (albeit most of the low hanging fruit has long since gone), which is why I said it's a mistake to *assume*.
No, it is not a mistake to assume. It is a mistake to present extreme edge cases as though they were equal to the overwhelming majority in importance. They are not.
There are quite a few businesses out there that couldn't exist without taking the Open Source route.
I haven't seen a business in 15 years that functioned without at least some open source software. What's your point? That doesn't mean that open source is the rational choice target for migrating the overwhelming majority of Windows-based workloads, or even most, or even more than "an insignificant number of edge cases". Especially with a time frame of one month. And the timeframe target is absolutely vital to the discussion in both articles.
It's not an Windows or Linux equation, Open Source has grown more through opening up new markets than cannibalizing Windows market share - obvious examples being bits of iOS & Android.
Why are you preaching here? Do you think that I don't know this? Or that I haven't said this? None of this has relevance to server workloads that are running on Windows 2003 Server that need to be migrated within the next month.
Don't proselytize irrelevant minutiae that has no bearing on the discussion to hand, it weakens your case...which was full enough of mindspiders as it is.
I am sure that I don't "belong" in the form of "IT" that you are espousing at the minute, and I am quite happy not to "belong" to it.
Thank $deity. I am espousing IT where practitioners make pragmatic judgements based on evident requirements, resources, time frames and return on investment. You clearly are having absolutely none of that. So do everyone a favour and stay the hell out of the industry before you destroy someone's company and with it all the jobs that company supports.
Rejoin the conversation when you can separate the evangelical message of open source you wish to proselytize from the cold hard reality of keeping businesses running. Then we can have rational discussions about which technologies to implement when and where and in what timeframes.
"If I upgraded a system without a prior change notice and it didn't work, I'd hope it was also the Print server, at least that way they couldn't print my P45 in the morning, as thats what would happen. I'm hoping the smilie face you added means you're being a little tongue in cheek :)"
If I upgraded a system without the proper change notice and it didn't work I'd roll back the VM/revert to the old VM. "Proper change notice" is a function of larger organizations. The larger the organization the more that matters, because the more people need to know.
When there are only two IT folk then "change notice" can be an IM or e-mail alerting about the change. It depends on what the change is, severity of anticipated impact, ease of reversion, etc.
Small organizations don't need to go through change assessment forms and so forth to determine what level of change notification is required. They also don't have to make change notification to more than the other nerds unless there's actually a good reason for someone else to know. It is a blessing.
If an application is "owned" by someone outside IT then I typically let them know that something is going to change (or has changed), just in case something odd happens that I didn't anticipate. The focus is more on the ability to smoothly roll back, however, than it is on carefully planning every change.
This is because the impact of most changes is small in small organizations. If we have to take a widget down for 15 minutes to revert it, the world doesn't end. I understand how that changes in enterprises and that's the reason for the smiley. I get away with things in smaller organizations that wouldn't fly in larger ones...but that doesn't mean that the things I'm doing are wrong for the situation. It means the risks and impacts are different and so where the time focus goes is different.
You live in a pretty black and white world, don't you?
Sure, some of the stuff I upgrade is simple enough to do overnight. A great example is a Win32 financials application that installs on a server and runs on Pervasive SQL. Porting that thing to a new Windows OS is as simple as
1) Install Pervasive SQL v.latest on Windows v.latest
2) Install application
3) Copy directory "backup batch" from live server and execute backup registration scripts
4) Put file named "GETOUT.NOW" in financials app directory "data" (this will boot any sessinos from people who didn't close the client)
5) Copy "data" to new server
6) Share "data" directory
7) Rename, readdress, reboot old VM
8) Verify client connectivity to new VM
9) Retire old VM
10) Don't bother manually doing any of the above because we scripted that in order to make reports servers automatically get a copy of the data from nightly backups about ten years ago
The above is a great example of a workload that migrates smoothly and seamlessly from one generation of Windows to the other. It doesn't cause fuss or muss and it doesn't migrate to *NIX worth a damn. It is also a very typical workload running on a Windows server.
I have zero problem bringing in people with expertise to let management know what the best upgrade path is. I don't care; I get paid the same whether the workload runs on Windows or Linux. The clients really don't care overmuch, excepting as it affects what they have to pay for things.
I have all sorts of problems with starting down the "consultant" bureaucratic fuckfest with a month left on the clock. Especially since I myself have enough experience to know what workloads are not going to go quietly into that good night and which ones might, with some coaxing.
The "professional" thing to do is to examine the business case for moving workloads from one platform to another and to be continually reevaluating the ROI of all deployed solutions. It is not to push a given religion, OS, etc on your employer or client.
It is absolutely unacceptable to attempt to use the eleventh hour of a refresh cycle to try to bring in massively disruptive changes. Doubly so if there exists the option to simply and smoothly move towards a supported solution with little or no disruption or downtime.
If the *NIX solutions are so amazing then they will be just as capable of making the case for their existence once the crisis has passed. if you need a crisis to exist in order for your solution to be viable then your solution is shit.
Sure. Samba 4 is good. Zentyal is good. Webmin is very good. Zarafa is also good for them as still run their own email. All stuff I use regularly. They do take a lot of learning though, and some - like Zarafa - are still very rough around the edges.
Still don't think you're going from a 2003 SBS to a Zentyal/Zarafa /w Samba4 & Webmin setup in a month unless you are amazing at migrations. And then...who maintains it? The guy who is a wizard at migrations and is probably out there making muchos money migrating everyone? Or the former Windows admins who have no idea what's going on?
"I still maintain a month isn't enough time to migrate from Windows"
This depends entirely on whether or not you tell anyone outside of IT that you're upgrading. I just tend to migrate and test over night and nobody asks questions. Other nerds are informed in case they have to support, but for the most part they just don't care, or are happy that another Server 2003 box has been binned.
There are advantages to not being bogged down by bureaucratic change management systems. :)
You're right re: shilling, my apologies.
I do however maintain that you are absolutely wrong re: "should declare MS partner" etc. It isn't remotely relevant, and hammering it really only serves to make you seem very much like one of the "OMG you're a shill" religious nutters.
"because you can totally migrate from 2003 to 2012R2 in a month"
Absolutely. I can do it in 45 minutes and complete testing. For applications that make the migration smoothly. (About 85% of applications in my experience.) For applications that don't, it can take anywhere from hours to years to solve the roadblocks.
...have you ever actually done migrations?
Funny, I average about three hours to move applications from Windows --> Windows, with full testing sweeps.
I average about 45 minutes to move OS-internal functions from Windows --> Windows with full testing sweeps.
Having your testing automated really helps. What, you mean you haven't automated testing? How the metric monkey fuck do you hope with patches?
Where migrations really cause problems is where there are client OS talking to server OS issues. (As opposed to client application talking to Win32 application running on a server). These can take months of testing to shake out the bugs.
Now. My testing will give me a "go/no go" return. If the case is "no go" - as, for example, an ongoing issue I have with Internet Explorer 8+ and printing - then there may not actually be fixes. This is why I still have applications living on Server 2003/Windows XP, Windows 2000 and even Windows NT.
Finding out why something doesn't work and then fixing it can take forever. But making sure that it does and will work should be very, very simple.
So migration is simple...and incredibly hard. All at the same time. Welcome to IT.
Not aimed at me, but all good questions. Let's answer.
How large is the largest network you ever managed?
25,000 (ish) nodes, (physical and virtual combined). Do BMCs count as nodes in and of themselves? They have to be updated...that would add another 5000.
How many servers (win and *nix), and what versions?
Complicated question. One...and many. One in that on my largest network everything was in fact driven by a single two-node lock-step cluster. This cluster handed out OSes over PXE to render nodes. The render nodes (5000ish in number) would pick up an operating system from the central cluster (from one of 4 different configs), boot from it and do work.
I have CentOS 5, CentOS 6, Windows XP and Server 2012 R2 all running as images. The nodes can be booted individually or in varying sized groups into each OS, which will give access to different rendering software to perform different tasks.
How many users and applications?
Well, the "biggest" network only has 8 users and 15 applications. My most users ever was about 750. they had about 1200 applications. My largest applications ever was about 1500 on a network with 250 users.
I'm sure most of those asserting you can migrate from OS A to OS B in a few days, never managed more than a few servers with some simplle, standard configurations, with a few standard apps and users - and never handled issues like special hardware support, fault tolerance, high availability, tiered storage, backups and so on... nor ever cared about security, really.
Not really a question, but...I've had to do all of that. Except usually on small enough budgets that I don't get to just ring up EMC for storage. I have to design and implement it myself. I've even build my own switches using Linux with realtime kernels and had to maintain every part of the stack from switches and NFV to storage to hypervisors to printers and endpoints.
And believe it or no, unlike most *nix worshipper, many MS users don't feel the need to justify their choice even against reason as it was a religion - if something deserve critics, they do, after all, it's just a technical choice, not a life one.
I feel I have to justify every single choice I make because everything costs money. If you can't justify it, why are you doing it?
Okay, so I've answered. Will the whingers please step up?
"Wait, what? Of course, that makes complete sense! That's what Microsoft partners do all the time - they are critical of Microsoft."
have you actually worked with Microsoft Partners? They basically bitch about Microsoft 24/7. Especially lately, since Microsoft drove down non-Azure margins and drove up SPLA pricing.
But hey, don't let reality intrude on your religion.
"I take it WINE is not a migration option for most of the Win32 applications?"
Wine's developers and community have spent their time working almost exclusively on end-user workloads. Desktop Linux stuff. Very - very - few server workloads are certified to run in Wine.
Also: how to you get stuff that's native to the Windows OS to "run in WINE"? Worse, there aren't like-for-like comparisons with the open source projects for all of them.
Consider SMB 3.0 with all of Microsoft's security features as an example. Samba isn't there yet - they certainly haven't added all the security integration in - so there is no direct like-for-like port. No autoscanning content to see if credit cards are being stored, no network access control integration, no integration with network egress defenses, etc.
You give up a fair amount of functionality if you move away from Microsoft's file system implementation. For my home network I might not care. My clients might have different needs on their own networks.
Even if you could, somehow, pull out the relevant bits from Microsoft's OS and run them in WINE...is that legal? Who provides support? How do you patch it?
A completely isolated server executable that could be run in WINE might be doable, but it's pretty rare these days. I've only found a handful that I could even get to function at all, and none that I'd consider viable to run in that state in perpetuity.
Patching and upgrades being the bitch: what happens if I get hit by a bus? Who can maintain that?
But most server executable aren't isolated. They rely on parts of Windows to work. (Many, for example, are increasingly taking advantage of the aforementioned security enhancements to Microsoft' file server stack.) Win32 applications get roots dug deep. The only way to change them over is usually to completely recode. In today's world that means making them into web applications. (Why recode for another OS just to get trapped there?)
You mean like every environment I've ever built? Yup. +1 for that.
It is a mistake to *assume* that Windows->Windows upgrades are always less effort than migrating off Windows
Who is assuming? I've done this for ages. I know what's difficult and what's not. NT --> 2000 was a pig. 2000 --> 2003 was not. 2003 --> 2008 was a pig for a very few edge cases, but 2003 --> 2008 R2 was not. 2003 --> 2012 or 2003 --> 2012 R2 can be a pig for some edge cases, but in the overwhleming majority of cases it is not.
NT --> 2000 (or later) and we could have a reasonable discussion about NT --> Linux being easier for up to 25% of workloads. 2000 --> anything later and I'm sorry, but moving to Linux is highly, highly unlikely to be easier. To the point that the number of instances in which Windows --> Linux is easier than Windows --> Windows fits within the margins of error for any testing methodology you care to name.
Which, quite frankly, is perfectly rational if you know anything about the technologies involved and have actually had to administer them in the field. It is absolutely a safe assumption to make that Windows --> Windows will be easier than Windows --> Linux with a one month timeframe remaining on the clock, because the number of instances where Windows --> Linux is easier than Windows --> Windows is irrelevantly small to start with.
If you can't accept that simple fact then you don't belong in IT. You should be out founding religions.
However, TP probably should have mentioned in the original article that his company is a Microsoft partner, that he profits through that partnership by obtaining free licenses, that he has "made a living from Microsoft" for "decades".
I've also made a living off Linux for decades. And off of BSD. And Solaris/OpenSolaris. And OSX. And Android (which I don't really consider to be Linux).
I have 8x to 12x as much *NIX deployed as Microsoft today (depending on how you count my clients). I think my company might also be a VMware partner, we've got some sort of affiliation with RedHat (but I can't remember what, exactly) and I've signed up for partner programs with ?????? other companies as well.
Being a "Microsoft Partner" doesn't mean I "profit through that partnership". It's a bunch of bureaucratic piss I have to go through in order to get what I need to do done. Just like I have an account with a Canada-local distributor to buy parts when clients need them. I don't charge the clients more than the markup for shipping and time spent assembling things. That's zero profit. It's just doing what needs to be done.
I make my money on retainers. Clients pay me per month. I give them best effort. That includes singing up to whatever ludicrous bullshit is required to buy things on behalf of my clients. if, sometimes, we can get a small break on licenses or something, great! If not, who cares? it determines what I install when I am setting up my own stuff and nothing more.
Which companies I've signed partnership agreements with isn't relevant. It isn't relevant because I'm not your typical MSP or VAR. I don't make my living selling what I'm told to sell. I don't make my living climbing up the ranks of their partner programs to get that extra 3% discount. I make exactly 0% profit on every MS license sold. I make exactly 0% profit on every server sold. I make exactly 0% profit on everything I sell to every single one of my clients because there is absolutely no point in dumping thousands of man-hours every year into partner programs to claw back 3 or 4 measly points on something.
Put simply: partnerships don't matter to me because I don't sell in volume, so I can't make any money off those partnerships. If you don't understand that, then you simply don't understand enough to be commenting on any of this.
Now, for the record, I absolutely do list anything I feel might potentially be of interest in determining how beholden I am to vendors. You can find that information here. Which would have taken about 10 seconds of Googling.
What's even funnier amidst your accusations of shilling and demands for "transparency" is that I more or less quit systems administration in January. I'm no longer a full time sysadmin. I maintain a stable of a handful of clients to ensure i keep my skills fresh, (there are only about 25 at current, and only two that are on monthly retainer,) but I make my money writing content now.
I have no interest in seeing Microsoft succeed, or in seeing them fail. Win or lose it doesn't affect me in any way.
Now quit trying to drown people in the name of rooting out witches and go outside and socialize with other human beings.
*shrug* I still have 12 Server 2003 --> 2008 Migrations to do this week. No big. Company has the 2008 licences siting around, and from experience, I'm 100% positive I can do each server in abou t 45 minutes.
Build parallel VM, install apps, migrate settings, change name on old VM to $name_old, change name on new VM to $name. Reboot, let DNS sort things out over night and Bob's your uncle.
Lots of Windows --> Windows migrations are just that easy.
All the Windows --> Linux stuff is done for this round. I think I've topped 250 servers this year.
I've done maaaaaybe 100 2003 --> 2012 R2 and another 20 2003 --> 2008. Over the past 5 years most 2003 upgrades have gone to CentOS for me, but most of that has been web servers or NFV.
Only a handful on Win32 apps ever made it to Linux. Win32 apps will persist for decades, I think, dragging Windows along, zombie-like for ages.
That's going to be a real pain when MS decides to go Midori. :/
"I think that most people here realize that the cloud is a little bit more complex than that basic definition."
I have no such illusions about the competence of commenttards. I might once have believed that, but I've recently become quite disillusioned in the basic level of technical understanding of the readership.
"Can somebody tell me the difference between Cloud and Thin Client?"
1) If you are talking about "Software as a Service" consumed through a browser (which is the main part of "cloud computing" that end users typically see) then the big difference is thin client versus browser.
Thin clients are thin. In the most traditional of thin clients (which we now term "zero clients" today) they have just enough brains to pass KVM back to the super and bring a screen to you. Modern "thin clients" are actually full "fat clients", but heavily locked down. That's a whole other discussion, as this is largely unnecessary, but done because it's easy.
Browsers are fat. They are basically their own operating systems at this point that run on top of another operating system. Browsers are the most common means to consume "cloud" apps, but not the only one. which leads us to number 2.
2) Cloud is about self service of IT resources. Those resources can be anything from a place to stick a fat, traditional Win32 application, to Desktop as a Service to a backup destination to a browser-based Software as a Service application.
The key differentiator between cloud-based and traditional architectures is that "cloud" removes people from the provisioning portion of the exercise. Developers can call resources through APIs. End users or administrators can create resources through a web interface. You don't fill out a request on a piece of paper, submit it to a bureaucracy and wait months.
Clouds can be internal to an organization (private cloud) consumed from an outsourcer (public cloud) or they can have the ability to use resources in either place (hybrid cloud).
That last is something that needs to be remembered. "Cloud" does not have to mean "the public cloud". You can build clouds in your own home if you want. Just start layering self-service interfaces on top of your existing network and voila: a private cloud.
"but given the quality of some of the devs I've worked with I'd be terrified at seeing them try to administer systems"
That could be said of any area of human endeavour though. Won't stop the inevitable. It'll just mean there's a whole lot of shitty DevOps out there and a handful of good ones. Plus ca change...
"I think to many people think of the cloud as putting up an Apache web server on a virtual box in a third third party datacenter somewhere."
Well, personally, I wouldn't call that a cloud. I'd call that a hosted solution. But that's just me.
"knowing how to tune applications to really make them sing is beyond a lot of sysadmins; you can change kernel params or registry settings etc. to help but without real in-depth knowledge you're limited in what you can do."
I think you make the point for Dev and Ops convergence quite clearly. The truth of the matter is that there are too many applications for traditional sysadmins to keep track of. You can't just have a generic "ops guy" and expect them to make everything run to plan.
That said, it's not that hard for a developer to learn about the environment they're developing for. DevOps who understand how the application they're coding interacts with the web server, the database, the kernel, etc will be better at creating efficient solutions than a traditional dev + ops arrangement.
Already, this is being noticed by companies who have adopted DevOps. IMHO, sheer economics will push devs towards taking on more and more of the ops role, leading to the traditional ops guys to evolve towards SecOps.
Excellent way to put it. 100% agreement.
...did you just imply that I am pro-cloud? If so, I'm getting your comment framed.
There's no haterade as delicious as developer haterade.
Apart from marketing, how is the cloud different to the network? What's a 'cloud solution' and how does that differ from two or more processes communicating over sockets?
The difference between a "cloud" solution and a non-cloud solution is that the cloud solution is self-service.
What is a "cloud" to the users of that solution is a network to the folks who have to stand it up. The lines can get blurry, but the difference is pretty simple to articulate. In a non-cloud network resources are allocated via manual intervention and approval after receiving a change request. In a cloud those seeking to consume network resources can go about consuming those resources without needing to submit change requests and get approval from IT.
That's the big difference.
ALL code is deterministic, even when pseudo random numbers are used. Its simply a case of understanding the algortihms and the code paths. Just because some code is highly complex doesn't make it non deterministic unless quantum CPUs have quietly snuck in without anyone noticing.
"Non-deterministic code" is a bit tongue in cheek. I am referring to the fact that the applications behave in a completely non-deterministic factor as far as systems administrators, end users, or other applications interfacing via API/XML Import/what-have-you are concerned.
Anywho, you may now continue with your rage-induced spleen venting. Enjoy.
So, you'll be letting the firemen set rates for putting out your house at their leisure, then? No taxes = no government = no emergency services. You ready to live in that world?
Maybe you should read up on the history of New York City.
Oh, yes. Sorry. I mean Microsoft PROBABLY isn't going to do that. So probably, in fact, that it bloody well certainly won't.
You are absolutely, completely and utterly delusional. Your brand tribalism is beyond hope of compensation. You are willing to trust a company - and not just any company, but Microsoft, who have done much to cause a loss of trust and nothing to earn it - with absolute faith. That's not only every single possible bad practice in risk management it's actually insane.
Yes, it is entirely possible that Microsoft will, in fact, force Windows 10 onto the world. Accidentally or otherwise. It may not be likely, but it is raw insanity to proclaim it as an impossibility.
Thus we have a possible event with a low probability but a potentially massive impact. The event can be guarded against with so minimal an input of time and resources as to be inconsequential. It is thus criminal negligence (in at least some jurisdictions) to ignore this possibility and do nothing to about it.
Whether I believe Microsoft will or won't do something is completely fucking irrelevant to the discussion. This discussion is about risk management practices in IT. Specifically it is about taking basic, pain-free steps to avoid possible calamity involving a vendor who is manifestly untrustworthy.
Personal opinions about a vendor should not enter into this at all. There is no room for them. It's a very simple risk management exercise.
As for why people feel the need to engage in risk management assessment and to be rather cynical about Microsoft's turstworthiness as part of those assessments, well...I'm going to leave you to ponder that on your own. Assuming you can overcome your brand tribalism enough to do it.
Oh, and next time you try to find some neat little cultural box to stuff me in so that you can malign me as part fof a group you have a religious hatred for, don't choose Linux. Presuming that I am somehow a Linux fanboy does nothing but demonstrate your overwhelming ignorance.
"Microsoft ISN'T going to install W10 on everyone's machine on D-day+1 as the Linux boys are suggesting. Jeez! Think what you're saying."
Microsoft most likely will not install W10 on everyone's machine on D-day +1. But you're a complete idiot if you rule out the possibility altogether. Not only that, D-day +1 is very limited.
Machines spontaneously installing Windows 10 on D-day +1 is probably a technical glitch. But the farther out we get from D-day+1 the more likely it is that some arrogant twunt will decide they know "what's best" and simply push it out as an auto update.
Now, that likelihood may be going from 0.0000000001% to 0.0000000002% with time (for etiher malice or incompetence,) but the possibility is emphatically not zero. There are examples in Microsoft's history of both malice and incompetence on the scale required to to something that egregious. Only a complete idiot would ignore the possibility.
Trust is earned. Microsoft have done nothing to earn it.
So you don't take the risk, no matter how small. You put the time and effort in to blocking the update as best you can. That way, if in the exceptionally unlikely event you are ever hauled in front of a judge to explain why you allowed something like this to occur (which caused license violations/citywide traffic jams/cats and dogs living together/whatever) you can say with absolute certainty you did everything you possibly could have to prevent it from occurring.
"I'm sorry, your honuor, but I trusted Microsoft not to screw up" won't fly.
If you cannot understand the above, please don't have anything more to do with computers, ever.
"What you're saying is once a criminal, always a criminal, right?"
No, I'm saying "screw me over once, shame on you. Screw me over twice, shame on me."
"The reality is that although Microsoft might get it wrong sometimes, they don't actually set out to do that"
I do not believe this to be true.
"There isn't some meeting somewhere going "oooh! How can we screw things up on purpose now?""
Yes there are. Several someones. I knew a few of them. For example: there are people whose job it is to screw over service providers so that Azure uptake will increase. There are people whose job it is to screw over those selling software so that Azure uptake will increase. There are people whose job it is to screw over small organizations so that they go elsewhere and Microsoft doesn't have to support them. And there are those whose job it is to determine the exact maximum that Microsoft can get away with licensing in areas where Microsoft has effective monopolies so as not to trigger scrutiny.
Microsoft employs many people whose sole job it is to screw us all over.
"All large companies - even... the ones who handle Linux... shock, horror - make mistakes. Screw up. P*** people off."
And? Do you think I am hugely enamored of Red "fuck you, bitches systemd ahahhahahahahahah" Hat? Or Canonical's "Amazon will destroy your sense of privacy because you suck ahahahahahahaha" Ubuntu? You're pretty funny.
"The simple fact is that the things that go "wrong" seem to stir up the anti-MS brigade far more than the normal population"
Wrong. The simple fact is that most people are far too forgiving of Microsoft's mistakes but tend to get really pissed when Microsoft does something deliberately assholeish. They should really be crucifying Microsoft for both.
"even to the point of said brigade predicting things will go wrong even before they've happened"
That's not a "brigade". That's people able to learn from history.
"In fact, the anti-MS sentiment even anticipates things will go wrong before they've been thought of!"
And yet, Microsoft rarely proves the cynics wrong.
"And that's precisely my original point."
That you don't actually understand how social dynamics works? Or that you are a raging brand tribalist who can't understand why people don't love your brand daddy and care about things like value for dollar, not getting screwed over and not being herded into choices they don't want by an arrogant company that doesn't care about them, their privacy or the security of their data?
Learn from history. Take a bit more cynical view of the companies to which you've attached your sense of self worth and you might not be constantly surprised by the companies in question or the reaction of everyday people to how they're treated by those large corporations.
Why would I use Windows Enterprise? It costs way more than Pro and requires a subscription. The subscription gets me nothing of any benefit and the total cost delta is insane. XP gave us 14 good years. 7 will give us 11 great years. In what universe is there value in buying Windows Enterprise?
Trust is earned, and Microsoft have done absolutely nothing to earn it.
In addition, real Windows administrators with actual experience in the field will remember many an instance of Windows downloading and installing updates despite both GPOs and local settings telling the damned thing not to. This causes all sorts of havoc when it reboots with a program running that doesn't autosave what it's doing, as so many older programs are prone to.
But sure, we're all crazy conspiracy theorists.
"licenses are payable to the government for £XX,XXXX.XX"
Not everyone uses a medallion system. In fact, globally, they're kind of rare. So no, in most jurisdictions it doesn't cost tens or hundreds of thousands to license a taxi. It costs you a commercial driver's license, commercial insurance and then you have to pay for inspections every X months at a licensed facility.
But hey, keep assuming that every single jurisdiction and every single economic influencer is the same everywhere. You'll go far that way.
Hitchhiking is illegal because it presents a public danger. Too many idiots trying to hitch on crowded 110 kph motorways, city ring roads, or during rush hour in the middle of the city. It caused accidents which in turn caused a lot of loss of life.
In Alberta we aren't as strict as our neighbours in British Columbia. But Alberta is (mostly) prairie and BC is (mostly) mountains. The highways are 100 kph in most parts of BC and if some dumb twat is pulled over on some bend where there's no shoulder (or on on of the tunnels) then the semi that's 500m behind him just might not be able to stop in time and very well might slam into him. Killing the idiot pulled over and his intended passenger.
In Alberta we don't have as many blind spots, but we absolutely do have "near, at or beyond" capacity major highways. When I say highway understand that I am not talking about "oh, it's a jaunty 30 minutes from some bedroom community to London." I'm talking about "it's 4 hours from Edmonton to Calgary" or "it's 6 hours from Edmonton to Ft Mac." I know truckers and riggers (usually parents going home to see thier kids) that do the whole 13.5 hour haul from Ft Mac to Lethbridge in a day, stopping only in Edmonton and Calgary for gas and to take a piss. this isn't rare, this is our whole goddamned province.
In Alberta, half the drivers are half asleep. They're driving along these great big long stretches of road at 120kph in the slow lane and 140kph in the fast lane on hiways that are 110kph limit. With the exception of "provincial NIMBY day" every three months (which is announced in the papers ahead of time) the cops don't pull anyone over for speeding, they flick on their lights and barrel down the hiway at 160kph to get to the next donut shop.
If you pull over on the side of the road you are dead.
If you go the speed limit (or lower) in the slow lane you'll be run off the road.
If you go the speed limit (or lower) in the fast lane, you'll be run over.
if you have a car problem on the highway, you drive it into the ditch and call the tow truck. We've got nearly 100% cell coverage so you just don't take the risk. The ditches are gentle slope and that's where non-functional cars belong.
Saskatchewan is twice as bad again, because it's even farther between major settlements and it's longer and straighter and flatter.
The cities aren't much better. A family of four might have 5 cars and an RV. One for each adult and teenager, plus a "hauler" (old pickup or candy van) and the RV. Everyone drives everywhere. Edmonton is a city of about a million people with a metro area the same size as London. the city proper - which is about half the size of Greater London, geographically - is functionally bumper-to-bumper traffic at 40kph for 6 hours a day, and not bumper-to-bumper at 70kph the rest of the day.
You don't pull over on the side of the road without backing up traffic for kilometers, potentially getting a ticket and at least getting scalding hot coffee thrown at you by motorists pulling around you.
Now, if you want to pull in to a gas station or something and pick up folk there, nobody is going to hassle you. If you find a roadside turnout or a rest stop (the provinces build rest stops so that people who are so exhausted they are willing to admit they're too tired to drive can park and sleep in their car about every 100km or so) then hey, go ahead and pick up hitchhikers.
But the whole notion of "stick your thumb out on the side of the highway and catch a ride" really only applies on the lightly used rural highways or in the smaller, not crazy-busy-filled-with-angry-bees towns.
Suggestions like "why doesn't everyone just drive slower" are going to be met with derision and laughter. Our provinces are huge, and they are sparsely populated. Our cities are spectacularly low density. There is no such thing as "living without a vehicle" here unless you happen to be willing to confine yourself to one of the "major" cities for almost your whole life. Even then, expect to take at least 1hr to get anywhere, probably 2 or 3, depending on if your journey is all bus or if you can shave some off but taking the LRT.
People get impatient after 30 minutes. Try driving for 4, 6 or 13.5 hours where there is nothing to see. It's just farms and cows and farms and canola and farms and canola and some more cows. The mind wanders. You go loopy. Everyone drives faster than they should.
BC is prettier, but after a while all mountains look alike and the need to get to civilization just so you can pee takes over. Maybe you're a good and courteous driver. Maybe him and her and them all are. But they won't all be. They won't all be awake, or not needing to pee, or paying attention, or courteous or whatever.
So that's hitchhiking here. It's dangerous. Hell, driving period is dangerous. But we've designed our infrastructure and laws to mostly accept the realities of things rather than trying to force our massively diverse populations (a significant % of people in all three provinces weren't born in those provinces, etc) to come to heel.
"It seems to me that Uber is basically a form of organised hitch-hiking "
Hitchhiking is illegal in many parts of the province in my province and both neighboring ones. So that still rules out Uber.
Bingo. Same here in Edmonton. Local council sets "maximum" rates (to which all cab companies align) and those rates must be displayed. This comes in the form of a sticker on the window. Cabs are calibrated regularly, inspected regularly, and all the drivers must have a commercial license and insurance. I believe - but am not 100% sure - that criminal records checks are enforced for cabbies too.
There is no medallion system here. Cab companies can field as many cabs as they want, but there is a limit to what's practicable. Anyone can start a cab company, but you'll be ruthlessly driven out of the market or bought by one of the big three if you try. (Though that was "the big two" 15 years ago...so entering the market is in fact possible.)
All the Taxi companies here have apps. In fact, the biggest one, which has several sub-"brands", has one for each brand. (Not ideal, I know.) They're all made by the same developer. *sigh*
Uber and Lyft are easier to use in that you can sit there and stare at your mobile and tap your feet and wait for the cab to arrive. But in every other way the local council-regulated taxies are better, sfare and more predictable.
"understanding the workload and suggesting/doing some "real world" like tests is part of doing the job right"
Wrong. It should be part of doing the job "right", but IT isn't a profession. There is no professional association and anyone can practice IT without needing to be government by a body with an ethics board. Thus "right" becomes less about "should" and instead becomes "whatever pays the mortgage with the least amount of effort".
The issue with "acceptable performance" is that it's hard to qualify with synthetics. You need to dupe actual workloads and then replay a day's worth of real world work on them. Personally, I like replay at 1.5x speed in order to allow for growth calculations, but that takes a lot of time and research beforehand to fully understand the workloads before testing.
I test with synthetics, but I find them almost meaningless. I far prefer to take the workloads I've babied every single day for the past 11 years and run them on $infrastucture. I know those workloads inside and out and i know how they respond to different types of kit. I can get a better feel for "good enough" or not using those workloads than any of the synthetics I run.
But that's me, and that's my workloads. Someone else is going to have a different mix. I think it's important we get as many different testers testing their real world mixes as possible so that we have a deep pool of knowledge available.
"Advertise my services"? To whom, precisely? Maxta? Maxta have me on speed dial. SimpliVIty? They know where to find me if they want me. Same for pretty much anyone else. I don't have to advertise. There aren't a lot of people writing technical content and there seems to be an unlimited demand for it.
If you think that my offering to settle the performance dispute for free is somehow "advertising" you're quite frankly insane. Do you have any idea - and idea at all - what it takes to do a full review on a hyperconverged solution? That's weeks of work, and VMware versus Nutanix is at least two units.
Maybe - maybe - I sell enough articles to make up my time, but not likely. And, to be entirely blunt about it, I could make a hell of a lot more money by basically rolling my face around on the keyboard and "writing what I know" instead of spending the time doing hard research.
Helping wingus and dingus grow up and stop cluttering everyone's twitter and news feeds doesn't really pay dividends except possibly as a portfolio item to say "here, look, I did this". But I have a Tintri review cooking, a SimpliVity one yet to write, I have a Scale versus Nodeweaver danceoff that's waiting and $diety knows what else. I have enough reviews that I could write until December and not need to look for more stuff to do.
That said, I want to see the end to this. Spy versus spy here is embarrassing, and it's clouding the issues. I end up taking a dozen calls a week now from people looking for the straight dirt on hyperconverged solutions, and half of them are trying to cut through the VMware versus Nutanix noise. Settling even one part of this loony tunes bullshit factory would earn me precious, precious hours of sleep.
As for the rest...
I don't know that Maxta have anything to lose by by open. You're probably correct in that Kiran and his lot are in a better position by being open than closed. But I can't agree with you on SimpliVity, or really, any of the others.
SimpliVity isn't the fastest of the lot. Full stop. They are the most consistent - that would be the accelerator card doing it's job - but they aren't the fastest. This is because Simplivity lays its blocks down on a RAID of magnetics and doesn't really cache to SSD quite as effectively as they should.
SimpliVity is absolutely open about this. They have no problems letting me - or anyone else - review the toys and publish the real world results. That buys them no real hoo-rah points except showing customers that they are honest and honourable. Which actually does count for a lot with some people.
Scale Computing? Scale doesn't even have flash! They certainly don't see benefit from letting me test their stuff. Yet Scale sent me some nodes, as well as to numerous other "thought leaders" and we've all had a right good go at them. We broadly agree on the pros and cons and so a picture of just what Scale is like from top to bottom has emerged.
That picture is not one that says Scale is the fastest, or the best priced, or the, well...best anything really. (Except possibly best support.) Scale has very specific tradeoffs and it seems rather a lot of companies are perfectly willing to accept those...and they're happy knowing right up front what the tradeoffs are.
On and on. Yottabyte, Nodeweaver, Tintri, Tegile, Nexenta, you name a storage company and - with a few childish exceptions - they're all open, helpful and friendly. Seems they understand that trust is hard to earn and easy to lose.
So here's an idea, mate: why don't you climb down off your high horse and learn a thing or two? The world is quite obviously far more complex than you imagine. Not to mention your terrible reading comprehension.
When you can conceive that people's motivations in the world extend beyond advertising and marketing and thwarting one's rivals maybe you'll be ready to play with adults. Maybe you'll even be person enough to use your real name, instead of lobbing accusations from behind the veil of an anonymous coward.
Until then, kindly get bent.
"El Reg readers will know what an ISO file is, but the average member of the public will not."
Windows Insiders should.
I betaed Windows 10 and all I got was an arrow to the knee.
"Smaller footprint than its predecessors"
"faster start-up and shutdown"
Who cares? Everyone has SSDs, and we only reboot once a month when updates force us to.
Only if application developers take advantage of it. In the meantime, Microsoft is making us all give up privacy in order to get this possible security.
"allows you to run Universal apps"
This is not a positive thing.
"and brings you into the 21st Century."
Funny, I've many an operating system that is 21st Century-enabled. Why is this Microsoft's first?
" the .NET framework (which underpins much of Exchange nowadays) allocates memory and CPU threads ineffectively"
Which is a fucking joke. Especially considering the hardware that's available today.
"But hey, the software's architects, the support teams who troubleshoot this stuff day in and day out, and the guys who have deployed that system for multiple millions of users - they don't know what they're talking about. But some VMware guy - I guess he must be an _expert_."
Yeah, you know, "some VMware guy" very well might be the expert here. Microsoft and it's developers and systems administrators don't need to care about money, or efficiency. They don't pay for the software licenses and they don't seem to give a bent damn about making the most use out of the hardware or datacenter space.
Put simply: Microsoft's priorities are clearly not the same priorities as actual businesses. So yes, I don't believe Microsoft are the experts here.
Here's an idea: cut Microsoft in half. Azure Public Cloud to be it's own thing, and "them who sell software" to be another. Now let's give it a year and see what the Azure teams have to say about the software after they start having to pay for it, and they start having to sustain and grow only on the backs of their own profits.