Category Archives: professional

.net, microsoft, professional, software development, visual studio

Visual Studio “14” announced, preview available with “Roslyn” open source compiler

Microsoft’s Soma Somasegar has announced the next version of Visual Studio, currently known as Visual Studio 14, but likely to be fully released in 2015 (and, I am guessing, likely to be called Visual Studio 2015).

This is a major release. It includes a new VB and C# compiler which is itself written in managed code, codenamed Roslyn. The open source Roslyn project provides new APIs that enable more powerful IDE features. Visual Basic is getting refactoring support for the first time.

The preview also includes a major update to ASP.NET that unifies ASP.NET MVC and the ASP.NET Web API, and has a new deployment model and developer experience:

Thanks to the Rosyln compiler, if you change ".cs” files or project.json file and want to see the change in the browser, you don’t need to build the project any more. Just refresh the browser.

…

There is no IIS express, nor IIS involved when you run from the command line. It means that you can publish your website to a USB drive, and run it by double clicking the web.cmd file!

On the C++ side, there is improved C++ 11 support and more features from C++ 14:

The Visual Studio "14" CTP includes support for user-defined literals, noexcept, alignof and alignas, and inheriting constructors from C++11, generalized lambda capture, auto function return type deduction, and generic lambdas from C++14, as well as many more new C++ features.

says Somasegar. There is also a refactored C Runtime (CRT):

msvcr140.dll no longer exists. It is replaced by a trio of DLLs: vcruntime140.dll, appcrt140.dll, and desktopcrt140.dll.

If you install the CTP (mine is downloading) use a spare machine or VM; it is an early preview that does not work side-by-side with other versions and the only uninstall may be to flatten the machine:

Installing a CTP release will place a computer in an unsupported state. For that reason, we recommend only installing CTP releases in a virtual machine, or on a computer that is available for reformatting.

apple, development, iphone, professional, software development

Apple’s Swift programming language: easy coding for OS X and iOS at last?

5 Comments

Apple has announced a new programming language, called Swift. (There was already a language called Swift, used for parallel scripting, but Apple links to the other Swift in case you land on the wrong page. So far it looks like the other Swift has not returned the favour).

For as long as I can remember, serious Apple developers have had to use Objective-C, an object-oriented C that is not like C++. I have only dabbled in Objective-C but when I last tried it I was pleasantly surprised: memory management was no hassle and I found it productive. Nevertheless it is an intimidating language if you come from a background of, say, JavaScript or Microsoft .NET. Apple’s focus on Objective-C has left a gap for easier to use alternatives, though the main reason developers use something other than Objective-C, as far as I am aware, is for cross-platform projects. Companies such as Xamarin and Embarcadero (with Delphi) have had some success, and of course Adobe PhoneGap (or the open source Cordova) has had significant take-up for cross-platform code based on HTML and JavaScript.

I should mention that RAD (Rapid Application Development) on OS X has long been possible using the wholly-owned Filemaker, a database manager with a powerful scripting language, but this is not suitable for general-purpose apps.

Overall, it is fair to say that coding for OS X and iOS has a higher bar than for Windows because Apple has not provided anything like Microsoft’s C# or Visual Basic, type-safe languages with easy form builders that let you snap together an application in a short time, while still being powerful enough for almost any purpose. This has been a differentiator for Windows. Visual Basic is almost as old as Windows itself, and C# was introduced in 2000.

Now Apple has come up with its own equivalent. I am new to Swift as are most people outside Apple, but took a quick look at the book, The Swift Programming Language, along with the announcement details. A few highlights:

  • Swift is a type-safe language that compiles to native code using LLVM.
  • The IDE for Swift is Xcode. It supports Cocoa development (Apple’s user interface framework) via import of the existing Objective-C frameworks, which become Swift APIs via the import keyword:

import UIKit

  • You can mix Swift and Objective-C in a single project. In Objective C you can use #import to make Swift code visible and usable.
  • Swift is a C-family language and you will find familiar features like curly braces and semi-colons to terminate lines (though semi-colons are optional).
  • Swift uses reference counting for automatic memory management. There is rather complex section in the book about weak references and unowned references, to solve some of the problems inherent in reference counting.
  • Type inference is the preferred approach to declaring the type of a variable, but you can state the type if required. You can also declare constants.
  • Swift supports single inheritance for classes and multiple inheritance for protocols (protocols are more or less equivalent to interfaces in other languages).
  • There are advanced features including closures, generics, tuples, and variadic parameters. (I am not sure if “advanced” is the right word, but other languages such as C# and Java took a while to get these).
  • Swift has something like destructors which it calls deinitializers.
  • There is an interesting feature called Extensions which lets you add methods to any existing type. For example, you could extend Int with a prettyprint method and then call 3.prettyprint.
  • Swift variables are not normally nullable; they must have a value. However you can declare optional types (add a ?, such as Int?) that can be set to nil. You can also declare implicitly unwrapped optionals which can be nil, but once assigned a value cannot be nil thereafter.
  • Swift includes the AnyObject type which can represent anything.

Swift seems to me to have similar goals to Microsoft’s C#: easier and safer than C or C++, but intended for any use right up to large and complex applications. One of the best things about it is the smooth interoperability with Objective-C; this also saves Apple from having to write native Swift frameworks for its entire stack.

A smart move? I think so, though Swift is different enough from any other language that developers have some learning to do.

What difference will Swift make? Initially, not that much. Objective-C developers now have a choice and some will move over or start mixing and matching, but Swift is still single-platform and will not change the developer landscape. That said, Swift may make Apple’s platform more attractive to business developers, for whom C# or Java is currently more productive; and perhaps Apple could find ways of using Swift in places where previously you would have to use AppleScript, extending its usefulness.

If Apple developers were tempted towards Xamarin or Delphi for productivity, as opposed to cross-platform, they will probably now use Swift; but I doubt there were all that many in that particular group.

I would be interested to hear from developers though: what do you think of Swift?

.net, android, apple, development, microsoft, mobile, professional, software development, visual studio, windows

Hands on with Xamarin 3.0: a cross-platform breakthrough for Visual Studio

Today Xamarin announced version 3.0 of its cross-platform mobile development tools, which let you target Android and iOS with C# and .NET. I have been trying a late beta preview.

In order to use Xamarin 3.0 with iOS support you do need a Mac. However, you can do essentially all of your development in Visual Studio, and just use the Mac for debugging.

To get started, I installed Xamarin 3.0 on both Windows (with Visual Studio 2013 installed) and on a Mac Mini on the same network.

image

Unfortunately I was not able to sit back and relax. I got an error installing Xamarin Studio, following which the installer would not proceed further. My solution was to download the full DMG (Mac virtual disk image) for Xamarin Studio and run that separately. This worked, and I was able to complete the install with the combined installer.

When you start a Visual Studio iOS project, you are prompted to pair with a Mac. To do this, you run a utility on the Mac called Xamarin.IOS Build Host, which generates a PIN. You enter the PIN in Visual Studio and then pairing is active.

image

Once paired, you can create or open iOS Storyboard projects in Visual Studio, and use Xamarin’s amazing visual designer.

image

Please click this image to open it full-size. What you are seeing is a native iOS Storyboard file open in Visual Studio 2013 and rendering the iOS controls. On the left is a palette of visual components I can add to the Storyboard. On the right is the normal Visual Studio solution explorer and property inspector.

The way this works, according to what Xamarin CEO Nat Friedman told me, is that the controls are rendered using the iOS simulator on the Mac, and then transmitted to the Windows designer. Thus, what you see is exactly what the simulator will render at runtime. Friedman says it is better than the Xcode designer.

“The way we do event handling is far more intuitive than Xcode. It supports the new iOS 7 auto-layout feature. It allows you to live preview custom controls. Instead of getting a grey rectangle you can see it live rendered inside the canvas. We use the iOS native format for Storyboard files so you can open existing Storyboard files and edit them.”

I made a trivial change to the project, configured the project to debug on the iOS simulator, and hit Start. On the Mac side, the app opened in the simulator. On the Windows side, I have breakpoint debugging.

image

Now, I will not pretend that everything ran smoothly in the short time I have had the preview. I have had problems with the pairing after switching projects in Visual Studio. I also had to quit and restart the iOS Simulator in order to get rendering working again. This is an amazing experience though, combining remote debugging with a visual designer on Visual Studio in Windows that remote-renders design-time controls.

Still, time to look at another key new feature in Xamarin 3: Xamarin Forms. This is none other than our old friend XAML, implemented for iOS and Android. The Mono team has some experience implementing XAML on Linux, thanks to the Moonlight project which did Silverlight on Linux, but this is rather different. Xamarin forms does not do any custom drawing, but wraps native controls. In other words, it like is the Eclipse SWT approach for Java, and not like the Swing approach which does its own drawing. This is keeping with Xamarin’s philosophy of keeping apps as native as possible, even though the very existence of a cross-platform GUI framework is something of a compromise.

I have not had long to play with this. I did create a new Xamarin Forms project, and copy a few lines of XAML from a sample into a shared XAML file. Note that Xamarin Forms uses Shared Projects in Visual Studio, the same approach used by Microsoft’s Universal Apps. However, Xamarin Forms apps are NOT Universal Apps, since they do not support Windows 8 (yet).

image 

In a Shared Project, you have some code that is shared, and other code that is target-specific. By default hardly any code is shared, but you can move code to the shared node, or create new items there. I created XamFormsExample.xaml in the shared node, and amended App.cs so that it loads automatically. Then I ran the project in the Android emulator.

image

I was also able to run this on iOS using the remote connection.

I noticed a few things about the XAML. The namespace is:

xmlns="http://xamarin.com/schemas/2014/forms"
xmlns:x="http://schemas.microsoft.com/winfx/2009/xaml"

I have not seen this before. Microsoft’s XAML always seems to have a “2006” namespace. For example, this is for a Universal App:

xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
xmlns:x=http://schemas.microsoft.com/winfx/2006/xaml

However, XAML 2009 does exist and apparently can be used in limited circumstances:

In WPF, you can use XAML 2009 features, but only for XAML that is not WPF markup-compiled. Markup-compiled XAML and the BAML form of XAML do not currently support the XAML 2009 language keywords and features.

It’s odd, because of course Xamarin’s XAML is cut-down compared to Microsoft’s XAML. That said, I am not sure of the exact specification of XAML in Xamarin Forms. I have a draft reference but it is incomplete. I am not sure that styles are supported, which would be a major omission. However you do get layout managers including AbsoluteLayout, Grid, RelativeLayout and StackLayout. You also get controls (called Views) including Button, DatePicker, Editor, Entry (single line editor), Image, Label, ListView, OpenGLView, ProgressBar, SearchBar, Slider, TableView and WebView.

Xamarin is not making any claims for compatibility in its XAML implementation. There is no visual designer, and you cannot port from existing XAML code. The commitment to wrapping native controls may limit prospects for compatibility. However, Friedman did say that Xamarin hopes to support Universal Apps, ie. to run on Windows 8 as well as Windows Phone, iOS and Android. He said:

I think it is the right strategy, and if it does take off, which I think it will, we will support it.

Friedman says the partnership with Microsoft (which begin in November 2013) is now close, and it would be reasonable to assume that greater compatibility with Microsoft XAML is a future goal. Note that Xamarin 3 also supports Portable Class Libraries, so on the non-visual side sharing code with Microsoft projects should be straightforward.

Personally I think both the Xamarin forms and the iOS visual designer (which, note, does NOT support Xamarin Forms) are significant features. The iOS designer matters because you can now do almost all of your cross-platform mobile development within Visual Studio, even if you want to follow the old Xamarin model of a different, native user interface for each platform; and Xamarin Forms because it enables a new level of code sharing for Xamarin projects, as well as making XAML into a GUI language that you can use across all the most popular platforms. Note that I do have reservations about XAML; but it does tick the boxes for scaling to multiple form factors and for enormous flexibility.

apple, development, microsoft, mobile, professional, software development

Hands on with Cordova in Visual Studio

At TechEd this week, Microsoft announced Apache Cordova support in Visual Studio 2013. A Cordova app is HTML and JavaScript wrapped as a native app, with support for multiple platforms including iOS and Android. It is the open source part of Adobe’s PhoneGap product. I downloaded the preview from here and took a quick look.

There is a long list of dependencies which the preview offers to install on your behalf:

image

and

image

The list includes the Java SDK, Google Chrome and Apple iTunes. The documentation explains that Java is required for the Android build process, Chrome is required to run the Ripple emulator (so you could choose not to install if you do not require Ripple), and iTunes is required for deploying an app to an iOS device, though a Mac is also required.

The license terms for both Chrome and iTunes are long and onerous, plus iTunes is on my list of applications not to install on Windows if you want it to run fast. Chrome is already installed on my PC, and I unchecked iTunes.

Next, I ran Visual Studio and selected a Multi-Device Hybrid App project (I guess “Cordova app” was rejected as being too short and simple).

image

An annoyance is that if you use the default project location, it is incompatible because of spaces in the path:

image

The project opened, and being impatient I immediately hit Run.

When you build, and debug using the default Ripple emulator (which runs in Chrome, hence the dependency), Visual Studio grabs a ton of dependencies.

image

and eventually the app runs:

image

or you can debug in the Android emulator:

image

A good start.

Microsoft has some sample projects for AngularJS, BackboneJS and WinJS. This last is intriguing since you could emulate the Windows Phone look and feel (or something like it) on Android on iOS, though it would look far from native.

The preview is not feature-complete. The only supported device targets are Android 4.x, IOS 6 and 7, Windows 8.x Store apps, and Windows Phone 8.x. Windows Phone debugging does not work in this preview.

microsoft, professional, software development

Brief reflections on 50 years of BASIC

2 Comments

Beginner’s All-Purpose Symbolic Code (BASIC) has turned fifty, as reported on The Reg and by Jack Schofield on ZDNet. A great moment in computer history, or would we have been better off without it?

My first computer (a Commodore PET) ran Basic from ROM, and without it you could do nothing, though developers bypassed it by using the POKE command to write low-level instructions into memory. The language is meant to be forgiving (as far as a computer language can be) and English-like, at the expense of being a little more verbose. It is case-insensitive and does not require braces or semi-colons to indicate blocks or lines of code, which makes programming look less intimidating for beginners.

I graduated onto an Atari ST, for which there was an excellent Basic implementation called GFA Basic, fast and capable. This was great for writing utilities, though, though serious programming tended to use one of several strong C compilers: Lattice C, Mark Williams C, HiSoft C come to mind.

Basic also had a role, even on the ST, as a macro language for applications. For example, the Superbase database manager used a version of Basic.

The company most strongly associated with Basic though is Microsoft. A version of Basic came with MS-DOS.

image

Microsoft also supported Basic for professional development. Microsoft Basic Professional Development System 7.x was a well-regarded development tool for business applications, though commercial shrink-wrap software tended to be written in C or C++.

That trend followed through to the Windows graphical environment. Visual Basic (VB), which made it easy to code Windows applications, was perhaps the most significant Basic release in terms of its impact, especially when it reached version 3.0 with full database support. Its popularity was such that many developers felt wounded when Microsoft discontinued Visual Basic 6.0, a direct successor, in favour of Visual Basic .NET which is something incompatible and different.

Further, VB 6.0 or something very like it lives on today, in the form of Visual Basic for Applications as found in all recent versions of Microsoft Office.

image

Despite this, Basic is in decline. Most of the professional developers I meet at events like Build use C# in preference to Visual Basic, there being little reason not to. C# is the premier language of .NET, and Visual Basic gets in the way if you want to keep up with latest .NET developments. Xamarin, which lets you code in .NET for iOS and Android, supports C# but not Visual Basic. Once you come to terms with semi-colons, braces and case-sensitivity, there is no real advantage to Visual Basic and C# is no more difficult.

I do see Visual Basic still used in education though, as well as by some developers who either prefer the language or are so used to it that they see no need to change; and to be fair, Xamarin aside, there is little if anything you can do in C# that you cannot also do in VB and the output is more or less the same.

The Roslyn project, which will be part of the next version of C# and probably in the next release of Visual Studio, lets you paste C# code as VB and vice versa.

Nevertheless, I believe we will see further decline in Basic usage, especially as it is little used outside Microsoft’s platform.

Would it have been better if Microsoft has not adopted Basic so wholeheartedly? There are some problems with Basic, though it is possible to write excellent code in Basic just as you can write poor code in C#, Python, C, or other more fashionable languages. Some issues:

  • Early versions of Basic encouraged badly structured programming with keywords like GOTO and GOSUB resulting in intricate loops that were hard to follow or debug.
  • Basic abstracts how software works to such an extent that you do not learn some important programming concepts such as pointers, addresses, memory allocation.
  • There is no natural progression from Basic to the C-like languages which dominate computing (C,C++,JavaScript,C#).
  • Visual Basic encourages developers to mix GUI code and business logic in the same files, as well as building user interfaces that tend not to scale well.
  • Small and declining professional use means that Basic is less useful than many other languages in the job market.

That said, Basic powers many excellent business applications as well as introducing many to the wonders of programming, and deserves our respect.

iphone, microsoft, mobile, professional, visual studio, windows

Microsoft completes Nokia acquisition: what now for Windows Phone?

Microsoft has completed its acquisition of Nokia today, a milestone in the turbulent story both of Nokia and of Windows Phone, which Nokia adopted in the hope of establishing a “third ecosystem” to challenge Apple iOS and Google Android.

Rumour has it that the Nokia acquisition was controversial within Microsoft and a large factor in the departure of Steve Ballmer as CEO. However, even if Microsoft took the view that an independent Nokia was better for Windows Phone, it faced the risk that market pressure would drive Nokia to Android and weaken the platform. The beginnings of that process may have been under way, with the launch of the Nokia X Android-but-not-Google range of phones, but we will never know, since Microsoft decided on acquisition.

image

How important has Nokia been for Windows Phone? In my view, life-saving. Before Nokia, there was no manufacturer nor operator which really cared about the platform, and it showed in lacklustre hardware and half-hearted marketing efforts. Nokia came up with the distinctive Lumia brand and style, added a decent mapping service, and with its focus on the PureView camera technology, gave enthusiasts a reason to take a close look at its devices. It also saw an opportunity at the low end, and created some great value devices that opened up a new market for the operating system.

There were some blunders (the original Lumia 800 suffered many faults and terrible battery life on launch) and Lumia did not grow fast enough to restore Nokia to health, but to my mind it was a good effort.

Today, the general opinion of Windows Phone is that it is a strong smartphone operating system but suffers from a lack of high-quality apps. Users have to put up with the fact that most app vendors feel they are done if they support iOS and Android; and if there is a Windows Phone version of their app, it is often poor. That is not a great position for Microsoft/Nokia to be in, but it could be worse. Blackberry 10, which is also a decent mobile operating system, has been all-but written off as a viable contender.

Microsoft is fortunate in that, unlike Blackberry, it can to some extent create its own ecosystem. Office 365, Bing, OneDrive, Nokia’s maps, Azure for developers needing a cloud back-end: taken together they form a viable alternative. In this respect, Microsoft actually has an advantage over Apple, which lacks this breadth of services.

I have been reading the latest Developer Economics report from Vision Mobile. It is a good example of the neutral perspective on Windows Phone, though you will find it somewhat inconsistent:

Windows Phone sales picked up significantly in Q3 2013, showing a 140% increase year-on-year, fuelled primarily by low-end device sales. According to Kantar, Windows Phone sales in the three months running to Oct 2013, reached double-digit figures in some Western European markets. While this is certainly a positive sign for Microsoft they will continue facing an uphill struggle, in an increasingly unfavourable race against the two runaway leaders, iOS and Android.

The report emphasises that iOS and Android have won the mobile OS wars, but says that there are signs of hope for Microsoft:

Windows Phone Developer Mindshare has finally moved upwards, following positive market signals in the last two quarters. As we have frequently highlighted in past reports, the developer intent has always been there, with Windows Phone figuring at the top of our Developer Intentshare chart, but needed positive market signs in order to convert this interest into Mindshare. While the 26% Developer Mindshare is still less than half of that for iOS, Microsoft can now claim that over a quarter of developers that target mobile platforms are now actively developing for Windows Phone. […]

As a latecomer to a mobile market dominated by strong network effects, establishing a credible footprint in mobile remains a formidable challenge
for Microsoft. We believe that Microsoft may be better served in the long-run by leveraging the Android ecosystem as the deployment platform for Office
and Server businesses which are still growing.

Microsoft is in fact supporting iOS and Android as clients for its cloud services, as noted again at yesterday’s financial webcast, where CEO Satya Nadella talked about a strategy that goes across “devices some ours, some not ours.” It is a bit of both though, and the company is not showing any signs of weakening its own mobile efforts.

In my view reports like that from Vision Mobile miss a couple of factors. One is that Windows and Windows Phone are converging. They already use the same OS kernel, and at the Build conference earlier this month Microsoft announced Universal Apps that will run on both, and the ability for developers to sell an app once and have users install on both phone and full Windows.

This means that the future of Windows Phone and that of Windows itself are closely bound together. Longer term, they will either both fade away, or both succeed.

Windows remains a huge business for Microsoft, despite the decline of the PC, especially in business. Microsoft’s problem though is that adoption of Windows 8 has been relatively weak, and that those who do use it, largely live in the desktop environment rather than running Store apps (of which Universal Apps are a variant).

Despite the dismal progress so far for the Store apps platform and ecosystem, I believe it should be taken seriously. On paper it has many advantages, not only for touch control, but also in deployment, security, roaming data driven by the cloud, and discoverability through the store. Isolation from the core operating system protects users against the things that destroy desktop Windows, like unwanted extras foisted on users who simply need to update Java or Flash.

At Build we saw not only Universal Apps, but also a preview of Office in the Windows Runtime (Store app) environment. We also saw a preview of Store apps running within a window in the desktop environment, solving the jarring transition between desktop and Store app environments that unsettles users. If Microsoft gets this right, both Windows Phone and Windows tablets will be substantially more attractive.

Microsoft also has the ability to bind Windows Phone into its enterprise device management environment, System Center and InTune. In Windows Phone 8.1 the device management and security features businesses need are much improved. More is still needed; but the company should be able to build integration points that make it attractive to business customers already using products such as Active Directory, Microsoft Office, Office 365, System Center or InTune.

Another factor is the strength of Visual Studio for developers, especially as Microsoft improves its integration with cloud services like Azure and Office 365. You can use C# everywhere from cloud or server to mobile client.

image

Cortana is sure that Windows Phone is the best; but check out the Bing ad.

What then is the future of Windows Phone? Uncertain, as ever; but if Microsoft pulls off a smooth Nokia acquisition – leaving in place the things that enabled the company to build the Lumia brand – and if it delivers on the promise we saw at Build, of a strong unified platform, then I expect market share to continue to grow. If it can climb to 10% or 15%, it will be on the map for vendors and the app problems will ease.

On the other hand, if Microsoft/Nokia means a return to the ineffective marketing and strategy we saw before Nokia adopted Windows Phone, then I expect Windows Phone to follow Blackberry into oblivion.

I am positive, but Microsoft needs to execute carefully and quickly to win market share for its mobile platform.

adobe, android, cordova, professional

A little bit of magic from Adobe PhoneGap: instant app preview on a device

Adobe has released an update to its PhoneGap tools (build mobile apps using HTML and Javascript) which enabled instant preview on a device. You don’t need to compile and deploy; it is enough simply to save your changes using a text editor, and the running app preview on the connected device will refresh.

I tried a quick demo on Windows 8. PhoneGap installation or update is easy since it uses the node.js package manager: open a node command line and type npm install –g phonegap.

I had to update the Android SDK to version 19, which was the usual tedious update using the Android SDK Manager, but also worked smoothly.

Then I created a PhoneGap app and ran it in the emulator. So far so goo.

Next, I went to an Android phone and installed the PhoneGap Developer App.

image

Back on the PC, I started the built-in web server that serves the app refreshes to the device. Simply navigate to the app folder in the node command line and type: phonegap serve

Then on the device I ran the PhoneGap Developer App, and connected to the PhoneGap development server (IP no and port). The app preview appeared on the phone.

I then used Notepad to change the app home page slightly. Saved. The app preview refreshed almost instantly.

image

Here it is on the device:

image

A cool feature, though with some limitations. See Raymond Camden’s post for a few more details.

Currently iOS and Android are supported, with Windows Phone coming soon.

development, microsoft, professional, software development, windows

Windows problems: new users cannot log in, SQL Server 2014 install fails

Two issues I have seen recently:

1. A Windows 7 laptop which belonged to a developer and was being passed on to a new user. However, although you could create the new user, you could not log in as that user. The error was “User Profile Service service failed the logon. User profile cannot be loaded.”

I narrowed the problem down to an “Access denied” error when trying to create the profile, but we decided to restore Windows to factory settings (using recovery tools) since that was probably a better approach for the new user anyway – you never know what a developer may have tweaked or installed in Windows!

2. SQL Server 2014 Database Engine installation failing on Windows 8. The reported error: Could not find the Database Engine startup handle.

Annoyingly, you cannot attempt a repair install if you get this, since repair cannot run if there has not been a successful install in the first place. However this error does not rollback the installation, so you have the feature installed but not working. You have to remove the SQL Server feature using control panel, then you can retry. I got this on a couple of machines, tried a few things but failed, so used fresh VMs instead.

Others have run into this and there is a solution, which applies to both problems. They are actually the same, since SQL Server 2014 creates a new profile in the default install.

This solution means resetting the permissions on c:\users\default so that they are replaced with permissions inherited from the parent folder. This solution works; but it is not perfect, since these are the wrong permissions for the folder (too loose). Someone has done the job of finding the correct permissions for Windows 7. Applying all these is arduous though, and things may possibly have changed with updates since. In a production environment it would be better to narrow down the exact permission that is wrong, or to do a system restore.

If you are happy to risk it, navigate to c:\users in Windows Explorer and find the Default folder. This is hidden by default; you can make it visible using the View Options in Explorer; on the View tab choose Show hidden files, folders and drives. Next, right-click the c:\users\Default folder and choose Properties, then the Security tab, then Advanced, then click Change Permissions. Check the box:

Replace all child object permission entries with inheritable permission entries from this object

Click OK – it does not take long – and now new users can log in, and the SQL Server 2014 setup runs OK.

image

The problem seems to be common so it is likely that either an official update or a commonly-used third-party install is breaking these permissions. I would be glad to know what it is; it would also be good if Microsoft would build into Windows a feature that would restore default permissions for a system folder like this one.

Update: It may be sufficient to delete a single file. See this thread where event logs lead to a sqmdata…sqm file that, when deleted, fixed the issue. Check the Windows Application log for event 1509.

.net, development, microsoft, professional

Hands on: SQL Server 2014 with data files in Azure Blob Storage

2 Comments

One intriguing new feature in Micrsosoft’s SQL Server 2014 is the ability to create or attach databases whose files are in Azure blog storage. This sounds like something that would not work at all well: why would you want a database engine to mount files located hundreds or thousands of miles away? However, the feature is apparently baked deeply into SQL Server, according to this white paper (which is essential reading if you want to know more):

SQL Server 2014 integration with Windows Azure blob storage occurs at a deep level, directly into the SQL Server Storage Engine; SQL Server Data Files in Windows Azure is more than a simple adapter mechanism built on top of an existing software layer.

· The Manager Layer includes a new component called XFCB Credential Manager, which manages the security credentials necessary to access the Windows Azure blob containers and provides the necessary security interface; secrets are maintained encrypted and secured in the SQL Server built-in security repository in the master system database.

· The File Control Layer contains a new object called XFCB, which is the Windows Azure extension to the file control block (FCB) used to manage I/O against each single SQL Server data or log file on the NTFS file system; it implements all the APIs that are required for I/O against Windows Azure blob storage.

· At the Storage Layer, the SQL Server I/O Manager is now able to natively generate REST API calls to Windows Azure blob storage with minimal overhead and great efficiency; in addition, this component can generate information about performance counters and extended events (xEvents).

It also seems that the main target usage is SQL Server running on Azure VMs in the same region as the blog storage, removing latency concerns, though the wording of the explanation is curious, implying almost that on-premise connection is supported but should not be:

Although it is theoretically possible and officially supported, using an on-premises SQL Server 2014 installation and database files in Windows Azure blob storage is not recommended due to high network latency, which would hurt performance; for this reason, the main target scenario for this white paper is SQL Server 2014 installed in Windows Azure Virtual Machines (IaaS). This scenario provides immediate benefits for performance, data movement and portability, data virtualization, high availability and disaster recovery, and scalability limits.

If you use blob storage in this way on an Azure VM, then I/O goes through the Virtual Network Driver, whereas an Azure data disk uses the Virtual Disk Driver. This nicety may be the main reason to consider the feature.

I tried both scenarios: on-premise and from an Azure VM. I had some difficulty getting started, despite this seemingly exhaustive tutorial. I followed it, I thought, to the letter, but got either the error:

Unable to open the physical file "https://myaccount.blob.core.windows.net/sqldata/azuredb.mdf". Operating system error 86: "86(The specified network password is not correct.)".

or else

CREATE FILE encountered operating system error 1117(The request could not be performed because of an I/O device error.) while attempting to open or create the physical file https://myaccount.blob.core.windows.net/sqldata/azuredb.mdf

The problem turned out to relate to the Shared Access Signature required. The supposedly exhaustive tutorial merely refers you to the CloudBlobContainer.GetSharedAccessSignature method in the Azure SDK and offers an incomplete code snippet. I wrote C# code for this and was able to generate a Shared Access Signature but it did not work (see above). I found myself in the depths of the Azure SDK, wondering if I should use version 2.1 or 3.0, and whether I should use Microsoft.WindowsAzure.StorageClient.CloudBlobClient or Microsoft.WindowsAzure.Storage.Blob.CloudBlobClient. The tutorial is also not clear about exactly which part of the Shared Access Signature you should store in the SQL Server Credential Manager; it is a multipart string separated by ampersands.

I have still not fully worked it out, but discovered the very helpful Azure Storage Explorer on CodePlex. If you follow the instructions in the white paper referenced above, and use the Azure Storage Explorer to generate the Shared Access Signature, then it works. The project is open source, so with a little effort it should be possible to find and document the exact requirements.

image

I tried creating and using a database from my on-premise SQL Server 2014 and I find the performance remarkably good, considering. There is no doubt some smart caching going on under the covers. Selecting 1000 rows from a table took 11 seconds the first time, and was instant the second time. It seems to me viable, on my brief look, though I am not sure why you would want to do this. However it is a good demonstration of how cloud and on-premise are coming ever-closer.

image

Running from an Azure VM in the same region is a different case, though I would suggest detailed and intensive testing before going into production.