Search Results

Search found 32260 results on 1291 pages for 'post request'.

Page 187/1291 | < Previous Page | 183 184 185 186 187 188 189 190 191 192 193 194  | Next Page >

  • The blocking nature of aggregates

    - by Rob Farley
    I wrote a post recently about how query tuning isn’t just about how quickly the query runs – that if you have something (such as SSIS) that is consuming your data (and probably introducing a bottleneck), then it might be more important to have a query which focuses on getting the first bit of data out. You can read that post here.  In particular, we looked at two operators that could be used to ensure that a query returns only Distinct rows. and The Sort operator pulls in all the data, sorts it (discarding duplicates), and then pushes out the remaining rows. The Hash Match operator performs a Hashing function on each row as it comes in, and then looks to see if it’s created a Hash it’s seen before. If not, it pushes the row out. The Sort method is quicker, but has to wait until it’s gathered all the data before it can do the sort, and therefore blocks the data flow. But that was my last post. This one’s a bit different. This post is going to look at how Aggregate functions work, which ties nicely into this month’s T-SQL Tuesday. I’ve frequently explained about the fact that DISTINCT and GROUP BY are essentially the same function, although DISTINCT is the poorer cousin because you have less control over it, and you can’t apply aggregate functions. Just like the operators used for Distinct, there are different flavours of Aggregate operators – coming in blocking and non-blocking varieties. The example I like to use to explain this is a pile of playing cards. If I’m handed a pile of cards and asked to count how many cards there are in each suit, it’s going to help if the cards are already ordered. Suppose I’m playing a game of Bridge, I can easily glance at my hand and count how many there are in each suit, because I keep the pile of cards in order. Moving from left to right, I could tell you I have four Hearts in my hand, even before I’ve got to the end. By telling you that I have four Hearts as soon as I know, I demonstrate the principle of a non-blocking operation. This is known as a Stream Aggregate operation. It requires input which is sorted by whichever columns the grouping is on, and it will release a row as soon as the group changes – when I encounter a Spade, I know I don’t have any more Hearts in my hand. Alternatively, if the pile of cards are not sorted, I won’t know how many Hearts I have until I’ve looked through all the cards. In fact, to count them, I basically need to put them into little piles, and when I’ve finished making all those piles, I can count how many there are in each. Because I don’t know any of the final numbers until I’ve seen all the cards, this is blocking. This performs the aggregate function using a Hash Match. Observant readers will remember this from my Distinct example. You might remember that my earlier Hash Match operation – used for Distinct Flow – wasn’t blocking. But this one is. They’re essentially doing a similar operation, applying a Hash function to some data and seeing if the set of values have been seen before, but before, it needs more information than the mere existence of a new set of values, it needs to consider how many of them there are. A lot is dependent here on whether the data coming out of the source is sorted or not, and this is largely determined by the indexes that are being used. If you look in the Properties of an Index Scan, you’ll be able to see whether the order of the data is required by the plan. A property called Ordered will demonstrate this. In this particular example, the second plan is significantly faster, but is dependent on having ordered data. In fact, if I force a Stream Aggregate on unordered data (which I’m doing by telling it to use a different index), a Sort operation is needed, which makes my plan a lot slower. This is all very straight-forward stuff, and information that most people are fully aware of. I’m sure you’ve all read my good friend Paul White (@sql_kiwi)’s post on how the Query Optimizer chooses which type of aggregate function to apply. But let’s take a look at SQL Server Integration Services. SSIS gives us a Aggregate transformation for use in Data Flow Tasks, but it’s described as Blocking. The definitive article on Performance Tuning SSIS uses Sort and Aggregate as examples of Blocking Transformations. I’ve just shown you that Aggregate operations used by the Query Optimizer are not always blocking, but that the SSIS Aggregate component is an example of a blocking transformation. But is it always the case? After all, there are plenty of SSIS Performance Tuning talks out there that describe the value of sorted data in Data Flow Tasks, describing the IsSorted property that can be set through the Advanced Editor of your Source component. And so I set about testing the Aggregate transformation in SSIS, to prove for sure whether providing Sorted data would let the Aggregate transform behave like a Stream Aggregate. (Of course, I knew the answer already, but it helps to be able to demonstrate these things). A query that will produce a million rows in order was in order. Let me rephrase. I used a query which produced the numbers from 1 to 1000000, in a single field, ordered. The IsSorted flag was set on the source output, with the only column as SortKey 1. Performing an Aggregate function over this (counting the number of rows per distinct number) should produce an additional column with 1 in it. If this were being done in T-SQL, the ordered data would allow a Stream Aggregate to be used. In fact, if the Query Optimizer saw that the field had a Unique Index on it, it would be able to skip the Aggregate function completely, and just insert the value 1. This is a shortcut I wouldn’t be expecting from SSIS, but certainly the Stream behaviour would be nice. Unfortunately, it’s not the case. As you can see from the screenshots above, the data is pouring into the Aggregate function, and not being released until all million rows have been seen. It’s not doing a Stream Aggregate at all. This is expected behaviour. (I put that in bold, because I want you to realise this.) An SSIS transformation is a piece of code that runs. It’s a physical operation. When you write T-SQL and ask for an aggregation to be done, it’s a logical operation. The physical operation is either a Stream Aggregate or a Hash Match. In SSIS, you’re telling the system that you want a generic Aggregation, that will have to work with whatever data is passed in. I’m not saying that it wouldn’t be possible to make a sometimes-blocking aggregation component in SSIS. A Custom Component could be created which could detect whether the SortKeys columns of the input matched the Grouping columns of the Aggregation, and either call the blocking code or the non-blocking code as appropriate. One day I’ll make one of those, and publish it on my blog. I’ve done it before with a Script Component, but as Script components are single-use, I was able to handle the data knowing everything about my data flow already. As per my previous post – there are a lot of aspects in which tuning SSIS and tuning execution plans use similar concepts. In both situations, it really helps to have a feel for what’s going on behind the scenes. Considering whether an operation is blocking or not is extremely relevant to performance, and that it’s not always obvious from the surface. In a future post, I’ll show the impact of blocking v non-blocking and synchronous v asynchronous components in SSIS, using some of LobsterPot’s Script Components and Custom Components as examples. When I get that sorted, I’ll make a Stream Aggregate component available for download.

    Read the article

  • The blocking nature of aggregates

    - by Rob Farley
    I wrote a post recently about how query tuning isn’t just about how quickly the query runs – that if you have something (such as SSIS) that is consuming your data (and probably introducing a bottleneck), then it might be more important to have a query which focuses on getting the first bit of data out. You can read that post here.  In particular, we looked at two operators that could be used to ensure that a query returns only Distinct rows. and The Sort operator pulls in all the data, sorts it (discarding duplicates), and then pushes out the remaining rows. The Hash Match operator performs a Hashing function on each row as it comes in, and then looks to see if it’s created a Hash it’s seen before. If not, it pushes the row out. The Sort method is quicker, but has to wait until it’s gathered all the data before it can do the sort, and therefore blocks the data flow. But that was my last post. This one’s a bit different. This post is going to look at how Aggregate functions work, which ties nicely into this month’s T-SQL Tuesday. I’ve frequently explained about the fact that DISTINCT and GROUP BY are essentially the same function, although DISTINCT is the poorer cousin because you have less control over it, and you can’t apply aggregate functions. Just like the operators used for Distinct, there are different flavours of Aggregate operators – coming in blocking and non-blocking varieties. The example I like to use to explain this is a pile of playing cards. If I’m handed a pile of cards and asked to count how many cards there are in each suit, it’s going to help if the cards are already ordered. Suppose I’m playing a game of Bridge, I can easily glance at my hand and count how many there are in each suit, because I keep the pile of cards in order. Moving from left to right, I could tell you I have four Hearts in my hand, even before I’ve got to the end. By telling you that I have four Hearts as soon as I know, I demonstrate the principle of a non-blocking operation. This is known as a Stream Aggregate operation. It requires input which is sorted by whichever columns the grouping is on, and it will release a row as soon as the group changes – when I encounter a Spade, I know I don’t have any more Hearts in my hand. Alternatively, if the pile of cards are not sorted, I won’t know how many Hearts I have until I’ve looked through all the cards. In fact, to count them, I basically need to put them into little piles, and when I’ve finished making all those piles, I can count how many there are in each. Because I don’t know any of the final numbers until I’ve seen all the cards, this is blocking. This performs the aggregate function using a Hash Match. Observant readers will remember this from my Distinct example. You might remember that my earlier Hash Match operation – used for Distinct Flow – wasn’t blocking. But this one is. They’re essentially doing a similar operation, applying a Hash function to some data and seeing if the set of values have been seen before, but before, it needs more information than the mere existence of a new set of values, it needs to consider how many of them there are. A lot is dependent here on whether the data coming out of the source is sorted or not, and this is largely determined by the indexes that are being used. If you look in the Properties of an Index Scan, you’ll be able to see whether the order of the data is required by the plan. A property called Ordered will demonstrate this. In this particular example, the second plan is significantly faster, but is dependent on having ordered data. In fact, if I force a Stream Aggregate on unordered data (which I’m doing by telling it to use a different index), a Sort operation is needed, which makes my plan a lot slower. This is all very straight-forward stuff, and information that most people are fully aware of. I’m sure you’ve all read my good friend Paul White (@sql_kiwi)’s post on how the Query Optimizer chooses which type of aggregate function to apply. But let’s take a look at SQL Server Integration Services. SSIS gives us a Aggregate transformation for use in Data Flow Tasks, but it’s described as Blocking. The definitive article on Performance Tuning SSIS uses Sort and Aggregate as examples of Blocking Transformations. I’ve just shown you that Aggregate operations used by the Query Optimizer are not always blocking, but that the SSIS Aggregate component is an example of a blocking transformation. But is it always the case? After all, there are plenty of SSIS Performance Tuning talks out there that describe the value of sorted data in Data Flow Tasks, describing the IsSorted property that can be set through the Advanced Editor of your Source component. And so I set about testing the Aggregate transformation in SSIS, to prove for sure whether providing Sorted data would let the Aggregate transform behave like a Stream Aggregate. (Of course, I knew the answer already, but it helps to be able to demonstrate these things). A query that will produce a million rows in order was in order. Let me rephrase. I used a query which produced the numbers from 1 to 1000000, in a single field, ordered. The IsSorted flag was set on the source output, with the only column as SortKey 1. Performing an Aggregate function over this (counting the number of rows per distinct number) should produce an additional column with 1 in it. If this were being done in T-SQL, the ordered data would allow a Stream Aggregate to be used. In fact, if the Query Optimizer saw that the field had a Unique Index on it, it would be able to skip the Aggregate function completely, and just insert the value 1. This is a shortcut I wouldn’t be expecting from SSIS, but certainly the Stream behaviour would be nice. Unfortunately, it’s not the case. As you can see from the screenshots above, the data is pouring into the Aggregate function, and not being released until all million rows have been seen. It’s not doing a Stream Aggregate at all. This is expected behaviour. (I put that in bold, because I want you to realise this.) An SSIS transformation is a piece of code that runs. It’s a physical operation. When you write T-SQL and ask for an aggregation to be done, it’s a logical operation. The physical operation is either a Stream Aggregate or a Hash Match. In SSIS, you’re telling the system that you want a generic Aggregation, that will have to work with whatever data is passed in. I’m not saying that it wouldn’t be possible to make a sometimes-blocking aggregation component in SSIS. A Custom Component could be created which could detect whether the SortKeys columns of the input matched the Grouping columns of the Aggregation, and either call the blocking code or the non-blocking code as appropriate. One day I’ll make one of those, and publish it on my blog. I’ve done it before with a Script Component, but as Script components are single-use, I was able to handle the data knowing everything about my data flow already. As per my previous post – there are a lot of aspects in which tuning SSIS and tuning execution plans use similar concepts. In both situations, it really helps to have a feel for what’s going on behind the scenes. Considering whether an operation is blocking or not is extremely relevant to performance, and that it’s not always obvious from the surface. In a future post, I’ll show the impact of blocking v non-blocking and synchronous v asynchronous components in SSIS, using some of LobsterPot’s Script Components and Custom Components as examples. When I get that sorted, I’ll make a Stream Aggregate component available for download.

    Read the article

  • Testing Workflows &ndash; Test-After

    - by Timothy Klenke
    Originally posted on: http://geekswithblogs.net/TimothyK/archive/2014/05/30/testing-workflows-ndash-test-after.aspxIn this post I’m going to outline a few common methods that can be used to increase the coverage of of your test suite.  This won’t be yet another post on why you should be doing testing; there are plenty of those types of posts already out there.  Assuming you know you should be testing, then comes the problem of how do I actual fit that into my day job.  When the opportunity to automate testing comes do you take it, or do you even recognize it? There are a lot of ways (workflows) to go about creating automated tests, just like there are many workflows to writing a program.  When writing a program you can do it from a top-down approach where you write the main skeleton of the algorithm and call out to dummy stub functions, or a bottom-up approach where the low level functionality is fully implement before it is quickly wired together at the end.  Both approaches are perfectly valid under certain contexts. Each approach you are skilled at applying is another tool in your tool belt.  The more vectors of attack you have on a problem – the better.  So here is a short, incomplete list of some of the workflows that can be applied to increasing the amount of automation in your testing and level of quality in general.  Think of each workflow as an opportunity that is available for you to take. Test workflows basically fall into 2 categories:  test first or test after.  Test first is the best approach.  However, this post isn’t about the one and only best approach.  I want to focus more on the lesser known, less ideal approaches that still provide an opportunity for adding tests.  In this post I’ll enumerate some test-after workflows.  In my next post I’ll cover test-first. Bug Reporting When someone calls you up or forwards you a email with a vague description of a bug its usually standard procedure to create or verify a reproduction plan for the bug via manual testing and log that in a bug tracking system.  This can be problematic.  Often reproduction plans when written down might skip a step that seemed obvious to the tester at the time or they might be missing some crucial environment setting. Instead of data entry into a bug tracking system, try opening up the test project and adding a failing unit test to prove the bug.  The test project guarantees that all aspects of the environment are setup properly and no steps are missing.  The language in the test project is much more precise than the English that goes into a bug tracking system. This workflow can easily be extended for Enhancement Requests as well as Bug Reporting. Exploratory Testing Exploratory testing comes in when you aren’t sure how the system will behave in a new scenario.  The scenario wasn’t planned for in the initial system requirements and there isn’t an existing test for it.  By definition the system behaviour is “undefined”. So write a new unit test to define that behaviour.  Add assertions to the tests to confirm your assumptions.  The new test becomes part of the living system specification that is kept up to date with the test suite. Examples This workflow is especially good when developing APIs.  When you are finally done your production API then comes the job of writing documentation on how to consume the API.  Good documentation will also include code examples.  Don’t let these code examples merely exist in some accompanying manual; implement them in a test suite. Example tests and documentation do not have to be created after the production API is complete.  It is best to write the example code (tests) as you go just before the production code. Smoke Tests Every system has a typical use case.  This represents the basic, core functionality of the system.  If this fails after an upgrade the end users will be hosed and they will be scratching their heads as to how it could be possible that an update got released with this core functionality broken. The tests for this core functionality are referred to as “smoke tests”.  It is a good idea to have them automated and run with each build in order to avoid extreme embarrassment and angry customers. Coverage Analysis Code coverage analysis is a tool that reports how much of the production code base is exercised by the test suite.  In Visual Studio this can be found under the Test main menu item. The tool will report a total number for the code coverage, which can be anywhere between 0 and 100%.  Coverage Analysis shouldn’t be used strictly for numbers reporting.  Companies shouldn’t set minimum coverage targets that mandate that all projects must have at least 80% or 100% test coverage.  These arbitrary requirements just invite gaming of the coverage analysis, which makes the numbers useless. The analysis tool will break down the coverage by the various classes and methods in projects.  Instead of focusing on the total number, drill down into this view and see which classes have high or low coverage.  It you are surprised by a low number on a class this is an opportunity to add tests. When drilling through the classes there will be generally two types of reaction to a surprising low test coverage number.  The first reaction type is a recognition that there is low hanging fruit to be picked.  There may be some classes or methods that aren’t being tested, which could easy be.  The other reaction type is “OMG”.  This were you find a critical piece of code that isn’t under test.  In both cases, go and add the missing tests. Test Refactoring The general theme of this post up to this point has been how to add more and more tests to a test suite.  I’ll step back from that a bit and remind that every line of code is a liability.  Each line of code has to be read and maintained, which costs money.  This is true regardless whether the code is production code or test code. Remember that the primary goal of the test suite is that it be easy to read so that people can easily determine the specifications of the system.  Make sure that adding more and more tests doesn’t interfere with this primary goal. Perform code reviews on the test suite as often as on production code.  Hold the test code up to the same high readability standards as the production code.  If the tests are hard to read then change them.  Look to remove duplication.  Duplicate setup code between two or more test methods that can be moved to a shared function.  Entire test methods can be removed if it is found that the scenario it tests is covered by other tests.  Its OK to delete a test that isn’t pulling its own weight anymore. Remember to only start refactoring when all the test are green.  Don’t refactor the tests and the production code at the same time.  An automated test suite can be thought of as a double entry book keeping system.  The unchanging, passing production code serves as the tests for the test suite while refactoring the tests. As with all refactoring, it is best to fit this into your regular work rather than asking for time later to get it done.  Fit this into the standard red-green-refactor cycle.  The refactor step no only applies to production code but also the tests, but not at the same time.  Perhaps the cycle should be called red-green-refactor production-refactor tests (not quite as catchy).   That about covers most of the test-after workflows I can think of.  In my next post I’ll get into test-first workflows.

    Read the article

  • Consuming the Amazon S3 service from a Win8 Metro Application

    - by cibrax
    As many of the existing Http APIs for Cloud Services, AWS also provides a set of different platform SDKs for hiding many of complexities present in the APIs. While there is a platform SDK for .NET, which is open source and available in C#, that SDK does not work in Win8 Metro Applications for the changes introduced in WinRT. WinRT offers a complete different set of APIs for doing I/O operations such as doing http calls or using cryptography for signing or encrypting data, two aspects that are absolutely necessary for consuming AWS. All the I/O APIs available as part of WinRT are asynchronous, and uses the TPL model for .NET applications (HTML and JavaScript Metro applications use a model based in promises, which is similar concept).  In the case of S3, the http Authorization header is used for two purposes, authenticating clients and make sure the messages were not altered while they were in transit. For doing that, it uses a signature or hash of the message content and some of the headers using a symmetric key (That's just one of the available mechanisms). Windows Azure for example also uses the same mechanism in many of its APIs. There are three challenges that any developer working for first time in Metro will have to face to consume S3, the new WinRT APIs, the asynchronous nature of them and the complexity introduced for generating the Authorization header. Having said that, I decided to write this post with some of the gotchas I found myself trying to consume this Amazon service. 1. Generating the signature for the Authorization header All the cryptography APIs in WinRT are available under Windows.Security.Cryptography namespace. Many of operations available in these APIs uses the concept of buffers (IBuffer) for representing a chunk of binary data. As you will see in the example below, these buffers are mainly generated with the use of static methods in a WinRT class CryptographicBuffer available as part of the namespace previously mentioned. private string DeriveAuthToken(string resource, string httpMethod, string timestamp) { var stringToSign = string.Format("{0}\n" + "\n" + "\n" + "\n" + "x-amz-date:{1}\n" + "/{2}/", httpMethod, timestamp, resource); var algorithm = MacAlgorithmProvider.OpenAlgorithm("HMAC_SHA1"); var keyMaterial = CryptographicBuffer.CreateFromByteArray(Encoding.UTF8.GetBytes(this.secret)); var hmacKey = algorithm.CreateKey(keyMaterial); var signature = CryptographicEngine.Sign( hmacKey, CryptographicBuffer.CreateFromByteArray(Encoding.UTF8.GetBytes(stringToSign)) ); return CryptographicBuffer.EncodeToBase64String(signature); } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } The algorithm that determines the information or content you need to use for generating the signature is very well described as part of the AWS documentation. In this case, this method is generating a signature required for creating a new bucket. A HmacSha1 hash is computed using a secret or symetric key provided by AWS in the management console. 2. Sending an Http Request to the S3 service WinRT also ships with the System.Net.Http.HttpClient that was first introduced some months ago with ASP.NET Web API. This client provides a rich interface on top the traditional WebHttpRequest class, and also solves some of limitations found in this last one. There are a few things that don't work with a raw WebHttpRequest such as setting the Host header, which is something absolutely required for consuming S3. Also, HttpClient is more friendly for doing unit tests, as it receives a HttpMessageHandler as part of the constructor that can fake to emulate a real http call. This is how the code for consuming the service with HttpClient looks like, public async Task<S3Response> CreateBucket(string name, string region = null, params string[] acl) { var timestamp = string.Format("{0:r}", DateTime.UtcNow); var auth = DeriveAuthToken(name, "PUT", timestamp); var request = new HttpRequestMessage(HttpMethod.Put, "http://s3.amazonaws.com/"); request.Headers.Host = string.Format("{0}.s3.amazonaws.com", name); request.Headers.TryAddWithoutValidation("Authorization", "AWS " + this.key + ":" + auth); request.Headers.Add("x-amz-date", timestamp); var client = new HttpClient(); var response = await client.SendAsync(request); return new S3Response { Succeed = response.StatusCode == HttpStatusCode.OK, Message = (response.Content != null) ? await response.Content.ReadAsStringAsync() : null }; } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } You will notice a few additional things in this code. By default, HttpClient validates the values for some well-know headers, and Authorization is one of them. It won't allow you to set a value with ":" on it, which is something that S3 expects. However, that's not a problem at all, as you can skip the validation by using the TryAddWithoutValidation method. Also, the code is heavily relying on the new async and await keywords to transform all the asynchronous calls into synchronous ones. In case you would want to unit test this code and faking the call to the real S3 service, you should have to modify it to inject a custom HttpMessageHandler into the HttpClient. The following implementation illustrates this concept, In case you would want to unit test this code and faking the call to the real S3 service, you should have to modify it to inject a custom HttpMessageHandler into the HttpClient. The following implementation illustrates this concept, public class FakeHttpMessageHandler : HttpMessageHandler { HttpResponseMessage response; public FakeHttpMessageHandler(HttpResponseMessage response) { this.response = response; } protected override Task<HttpResponseMessage> SendAsync(HttpRequestMessage request, System.Threading.CancellationToken cancellationToken) { var tcs = new TaskCompletionSource<HttpResponseMessage>(); tcs.SetResult(response); return tcs.Task; } } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } You can use this handler for injecting any response while you are unit testing the code.

    Read the article

  • Consuming the Amazon S3 service from a Win8 Metro Application

    - by cibrax
    As many of the existing Http APIs for Cloud Services, AWS also provides a set of different platform SDKs for hiding many of complexities present in the APIs. While there is a platform SDK for .NET, which is open source and available in C#, that SDK does not work in Win8 Metro Applications for the changes introduced in WinRT. WinRT offers a complete different set of APIs for doing I/O operations such as doing http calls or using cryptography for signing or encrypting data, two aspects that are absolutely necessary for consuming AWS. All the I/O APIs available as part of WinRT are asynchronous, and uses the TPL model for .NET applications (HTML and JavaScript Metro applications use a model based in promises, which is similar concept).  In the case of S3, the http Authorization header is used for two purposes, authenticating clients and make sure the messages were not altered while they were in transit. For doing that, it uses a signature or hash of the message content and some of the headers using a symmetric key (That's just one of the available mechanisms). Windows Azure for example also uses the same mechanism in many of its APIs. There are three challenges that any developer working for first time in Metro will have to face to consume S3, the new WinRT APIs, the asynchronous nature of them and the complexity introduced for generating the Authorization header. Having said that, I decided to write this post with some of the gotchas I found myself trying to consume this Amazon service. 1. Generating the signature for the Authorization header All the cryptography APIs in WinRT are available under Windows.Security.Cryptography namespace. Many of operations available in these APIs uses the concept of buffers (IBuffer) for representing a chunk of binary data. As you will see in the example below, these buffers are mainly generated with the use of static methods in a WinRT class CryptographicBuffer available as part of the namespace previously mentioned. private string DeriveAuthToken(string resource, string httpMethod, string timestamp) { var stringToSign = string.Format("{0}\n" + "\n" + "\n" + "\n" + "x-amz-date:{1}\n" + "/{2}/", httpMethod, timestamp, resource); var algorithm = MacAlgorithmProvider.OpenAlgorithm("HMAC_SHA1"); var keyMaterial = CryptographicBuffer.CreateFromByteArray(Encoding.UTF8.GetBytes(this.secret)); var hmacKey = algorithm.CreateKey(keyMaterial); var signature = CryptographicEngine.Sign( hmacKey, CryptographicBuffer.CreateFromByteArray(Encoding.UTF8.GetBytes(stringToSign)) ); return CryptographicBuffer.EncodeToBase64String(signature); } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } The algorithm that determines the information or content you need to use for generating the signature is very well described as part of the AWS documentation. In this case, this method is generating a signature required for creating a new bucket. A HmacSha1 hash is computed using a secret or symetric key provided by AWS in the management console. 2. Sending an Http Request to the S3 service WinRT also ships with the System.Net.Http.HttpClient that was first introduced some months ago with ASP.NET Web API. This client provides a rich interface on top the traditional WebHttpRequest class, and also solves some of limitations found in this last one. There are a few things that don't work with a raw WebHttpRequest such as setting the Host header, which is something absolutely required for consuming S3. Also, HttpClient is more friendly for doing unit tests, as it receives a HttpMessageHandler as part of the constructor that can fake to emulate a real http call. This is how the code for consuming the service with HttpClient looks like, public async Task<S3Response> CreateBucket(string name, string region = null, params string[] acl) { var timestamp = string.Format("{0:r}", DateTime.UtcNow); var auth = DeriveAuthToken(name, "PUT", timestamp); var request = new HttpRequestMessage(HttpMethod.Put, "http://s3.amazonaws.com/"); request.Headers.Host = string.Format("{0}.s3.amazonaws.com", name); request.Headers.TryAddWithoutValidation("Authorization", "AWS " + this.key + ":" + auth); request.Headers.Add("x-amz-date", timestamp); var client = new HttpClient(); var response = await client.SendAsync(request); return new S3Response { Succeed = response.StatusCode == HttpStatusCode.OK, Message = (response.Content != null) ? await response.Content.ReadAsStringAsync() : null }; } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } You will notice a few additional things in this code. By default, HttpClient validates the values for some well-know headers, and Authorization is one of them. It won't allow you to set a value with ":" on it, which is something that S3 expects. However, that's not a problem at all, as you can skip the validation by using the TryAddWithoutValidation method. Also, the code is heavily relying on the new async and await keywords to transform all the asynchronous calls into synchronous ones. In case you would want to unit test this code and faking the call to the real S3 service, you should have to modify it to inject a custom HttpMessageHandler into the HttpClient. The following implementation illustrates this concept, In case you would want to unit test this code and faking the call to the real S3 service, you should have to modify it to inject a custom HttpMessageHandler into the HttpClient. The following implementation illustrates this concept, public class FakeHttpMessageHandler : HttpMessageHandler { HttpResponseMessage response; public FakeHttpMessageHandler(HttpResponseMessage response) { this.response = response; } protected override Task<HttpResponseMessage> SendAsync(HttpRequestMessage request, System.Threading.CancellationToken cancellationToken) { var tcs = new TaskCompletionSource<HttpResponseMessage>(); tcs.SetResult(response); return tcs.Task; } } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } You can use this handler for injecting any response while you are unit testing the code.

    Read the article

  • Persisting Session Between Different Browser Instances

    - by imran_ku07
        Introduction:          By default inproc session's identifier cookie is saved in browser memory. This cookie is known as non persistent cookie identifier. This simply means that if the user closes his browser then the cookie is immediately removed. On the other hand cookies which stored on the user’s hard drive and can be reused for later visits are called persistent cookies. Persistent cookies are less used than nonpersistent cookies because of security. Simply because nonpersistent cookies makes session hijacking attacks more difficult and more limited. If you are using shared computer then there are lot of chances that your persistent session will be used by other shared members. However this is not always the case, lot of users desired that their session will remain persisted even they open two instances of same browser or when they close and open a new browser. So in this article i will provide a very simple way to persist your session even the browser is closed.   Description:          Let's create a simple ASP.NET Web Application. In this article i will use Web Form but it also works in MVC. Open Default.aspx.cs and add the following code in Page_Load.    protected void Page_Load(object sender, EventArgs e)        {            if (Session["Message"] != null)                Response.Write(Session["Message"].ToString());            Session["Message"] = "Hello, Imran";        }          This page simply shows a message if a session exist previously and set the session.          Now just run the application, you will just see an empty page on first try. After refreshing the page you will see the Message "Hello, Imran". Now just close the browser and reopen it or just open another browser instance, you will get the exactly same behavior when you run your application first time . Why the session is not persisted between browser instances. The simple reason is non persistent session cookie identifier. The session cookie identifier is not shared between browser instances. Now let's make it persistent.          To make your application share session between different browser instances just add the following code in global.asax.    protected void Application_PostMapRequestHandler(object sender, EventArgs e)           {               if (Request.Cookies["ASP.NET_SessionIdTemp"] != null)               {                   if (Request.Cookies["ASP.NET_SessionId"] == null)                       Request.Cookies.Add(new HttpCookie("ASP.NET_SessionId", Request.Cookies["ASP.NET_SessionIdTemp"].Value));                   else                       Request.Cookies["ASP.NET_SessionId"].Value = Request.Cookies["ASP.NET_SessionIdTemp"].Value;               }           }          protected void Application_PostRequestHandlerExecute(object sender, EventArgs e)        {             HttpCookie cookie = new HttpCookie("ASP.NET_SessionIdTemp", Session.SessionID);               cookie.Expires = DateTime.Now.AddMinutes(Session.Timeout);               Response.Cookies.Add(cookie);         }          This code simply state that during Application_PostRequestHandlerExecute(which is executed after HttpHandler) just add a persistent cookie ASP.NET_SessionIdTemp which contains the value of current user SessionID and sets the timeout to current user session timeout.          In Application_PostMapRequestHandler(which is executed just before th session is restored) we just check whether the Request cookie contains ASP.NET_SessionIdTemp. If yes then just add or update ASP.NET_SessionId cookie with ASP.NET_SessionIdTemp. So when a new browser instance is open, then a check will made that if ASP.NET_SessionIdTemp exist then simply add or update ASP.NET_SessionId cookie with ASP.NET_SessionIdTemp.          So run your application again, you will get the last closed browser session(if it is not expired).   Summary:          Persistence session is great way to increase the user usability. But always beware the security before doing this. However there are some cases in which you might need persistence session. In this article i just go through how to do this simply. So hopefully you will again enjoy this simple article too.

    Read the article

  • Introduction to Human Workflow 11g

    - by agiovannetti
    Human Workflow is a component of SOA Suite just like BPEL, Mediator, Business Rules, etc. The Human Workflow component allows you to incorporate human intervention in a business process. You can use Human Workflow to create a business process that requires a manager to approve purchase orders greater than $10,000; or a business process that handles article reviews in which a group of reviewers need to vote/approve an article before it gets published. Human Workflow can handle the task assignment and routing as well as the generation of notifications to the participants. There are three common patterns or usages of Human Workflow: 1) Approval Scenarios: manage documents and other transactional data through approval chains . For example: approve expense report, vacation approval, hiring approval, etc. 2) Reviews by multiple users or groups: group collaboration and review of documents or proposals. For example, processing a sales quote which is subject to review by multiple people. 3) Case Management: workflows around work management or case management. For example, processing a service request. This could be routed to various people who all need to modify the task. It may also incorporate ad hoc routing which is unknown at design time. SOA 11g Human Workflow includes the following features: Assignment and routing of tasks to the correct users or groups. Deadlines, escalations, notifications, and other features required for ensuring the timely performance of a task. Presentation of tasks to end users through a variety of mechanisms, including a Worklist application. Organization, filtering, prioritization and other features required for end users to productively perform their tasks. Reports, reassignments, load balancing and other features required by supervisors and business owners to manage the performance of tasks. Human Workflow Architecture The Human Workflow component is divided into 3 modules: the service interface, the task definition and the client interface module. The Service Interface handles the interaction with BPEL and other components. The Client Interface handles the presentation of task data through clients like the Worklist application, portals and notification channels. The task definition module is in charge of managing the lifecycle of a task. Who should get the task assigned? What should happen next with the task? When must the task be completed? Should the task be escalated?, etc Stages and Participants When you create a Human Task you need to specify how the task is assigned and routed. The first step is to define the stages and participants. A stage is just a logical group. A participant can be a user, a group of users or an application role. The participants indicate the type of assignment and routing that will be performed. Stages can be sequential or in parallel. You can combine them to create any usage you require. See diagram below: Assignment and Routing There are different ways a task can be assigned and routed: Single Approver: task is assigned to a single user, group or role. For example, a vacation request is assigned to a manager. If the manager approves or rejects the request, the employee is notified with the decision. If the task is assigned to a group then once one of managers acts on it, the task is completed. Parallel : task is assigned to a set of people that must work in parallel. This is commonly used for voting. For example, a task gets approved once 50% of the participants approve it. You can also set it up to be a unanimous vote. Serial : participants must work in sequence. The most common scenario for this is management chain escalation. FYI (For Your Information) : task is assigned to participants who can view it, add comments and attachments, but can not modify or complete the task. Task Actions The following is the list of actions that can be performed on a task: Claim : if a task is assigned to a group or multiple users, then the task must be claimed first to be able to act on it. Escalate : if the participant is not able to complete a task, he/she can escalate it. The task is reassigned to his/her manager (up one level in a hierarchy). Pushback : the task is sent back to the previous assignee. Reassign :if the participant is a manager, he/she can delegate a task to his/her reports. Release : if a task is assigned to a group or multiple users, it can be released if the user who claimed the task cannot complete the task. Any of the other assignees can claim and complete the task. Request Information and Submit Information : use when the participant needs to supply more information or to request more information from the task creator or any of the previous assignees. Suspend and Resume :if a task is not relevant, it can be suspended. A suspension is indefinite. It does not expire until Resume is used to resume working on the task. Withdraw : if the creator of a task does not want to continue with it, for example, he wants to cancel a vacation request, he can withdraw the task. The business process determines what happens next. Renew : if a task is about to expire, the participant can renew it. The task expiration date is extended one week. Notifications Human Workflow provides a mechanism for sending notifications to participants to alert them of changes on a task. Notifications can be sent via email, telephone voice message, instant messaging (IM) or short message service (SMS). Notifications can be sent when the task status changes to any of the following: Assigned/renewed/delegated/reassigned/escalated Completed Error Expired Request Info Resume Suspended Added/Updated comments and/or attachments Updated Outcome Withdraw Other Actions (e.g. acquiring a task) Here is an example of an email notification: Worklist Application Oracle BPM Worklist application is the default user interface included in SOA Suite. It allows users to access and act on tasks that have been assigned to them. For example, from the Worklist application, a loan agent can review loan applications or a manager can approve employee vacation requests. Through the Worklist Application users can: Perform authorized actions on tasks, acquire and check out shared tasks, define personal to-do tasks and define subtasks. Filter tasks view based on various criteria. Work with standard work queues, such as high priority tasks, tasks due soon and so on. Work queues allow users to create a custom view to group a subset of tasks in the worklist, for example, high priority tasks, tasks due in 24 hours, expense approval tasks and more. Define custom work queues. Gain proxy access to part of another user's tasks. Define custom vacation rules and delegation rules. Enable group owners to define task dispatching rules for shared tasks. Collect a complete workflow history and audit trail. Use digital signatures for tasks. Run reports like Unattended tasks, Tasks productivity, etc. Here is a screenshoot of what the Worklist Application looks like. On the right hand side you can see the tasks that have been assigned to the user and the task's detail. References Introduction to SOA Suite 11g Human Workflow Webcast Note 1452937.2 Human Workflow Information Center Using the Human Workflow Service Component 11.1.1.6 Human Workflow Samples Human Workflow APIs Java Docs

    Read the article

  • Help getting frame rate (fps) up in Python + Pygame

    - by Jordan Magnuson
    I am working on a little card-swapping world-travel game that I sort of envision as a cross between Bejeweled and the 10 Days geography board games. So far the coding has been going okay, but the frame rate is pretty bad... currently I'm getting low 20's on my Core 2 Duo. This is a problem since I'm creating the game for Intel's March developer competition, which is squarely aimed at netbooks packing underpowered Atom processors. Here's a screen from the game: ![www.necessarygames.com/my_games/betraveled/betraveled-fps.png][1] I am very new to Python and Pygame (this is the first thing I've used them for), and am sadly lacking in formal CS training... which is to say that I think there are probably A LOT of bad practices going on in my code, and A LOT that could be optimized. If some of you older Python hands wouldn't mind taking a look at my code and seeing if you can't find any obvious areas for optimization, I would be extremely grateful. You can download the full source code here: http://www.necessarygames.com/my_games/betraveled/betraveled_src0328.zip Compiled exe here: www.necessarygames.com/my_games/betraveled/betraveled_src0328.zip One thing I am concerned about is my event manager, which I feel may have some performance wholes in it, and another thing is my rendering... I'm pretty much just blitting everything to the screen all the time (see the render routines in my game_components.py below); I recently found out that you should only update the areas of the screen that have changed, but I'm still foggy on how that accomplished exactly... could this be a huge performance issue? Any thoughts are much appreciated! As usual, I'm happy to "tip" you for your time and energy via PayPal. Jordan Here are some bits of the source: Main.py #Remote imports import pygame from pygame.locals import * #Local imports import config import rooms from event_manager import * from events import * class RoomController(object): """Controls which room is currently active (eg Title Screen)""" def __init__(self, screen, ev_manager): self.room = None self.screen = screen self.ev_manager = ev_manager self.ev_manager.register_listener(self) self.room = self.set_room(config.room) def set_room(self, room_const): #Unregister old room from ev_manager if self.room: self.room.ev_manager.unregister_listener(self.room) self.room = None #Set new room based on const if room_const == config.TITLE_SCREEN: return rooms.TitleScreen(self.screen, self.ev_manager) elif room_const == config.GAME_MODE_ROOM: return rooms.GameModeRoom(self.screen, self.ev_manager) elif room_const == config.GAME_ROOM: return rooms.GameRoom(self.screen, self.ev_manager) elif room_const == config.HIGH_SCORES_ROOM: return rooms.HighScoresRoom(self.screen, self.ev_manager) def notify(self, event): if isinstance(event, ChangeRoomRequest): if event.game_mode: config.game_mode = event.game_mode self.room = self.set_room(event.new_room) def render(self, surface): self.room.render(surface) #Run game def main(): pygame.init() screen = pygame.display.set_mode(config.screen_size) ev_manager = EventManager() spinner = CPUSpinnerController(ev_manager) room_controller = RoomController(screen, ev_manager) pygame_event_controller = PyGameEventController(ev_manager) spinner.run() # this runs the main function if this script is called to run. # If it is imported as a module, we don't run the main function. if __name__ == "__main__": main() event_manager.py #Remote imports import pygame from pygame.locals import * #Local imports import config from events import * def debug( msg ): print "Debug Message: " + str(msg) class EventManager: #This object is responsible for coordinating most communication #between the Model, View, and Controller. def __init__(self): from weakref import WeakKeyDictionary self.listeners = WeakKeyDictionary() self.eventQueue= [] self.gui_app = None #---------------------------------------------------------------------- def register_listener(self, listener): self.listeners[listener] = 1 #---------------------------------------------------------------------- def unregister_listener(self, listener): if listener in self.listeners: del self.listeners[listener] #---------------------------------------------------------------------- def post(self, event): if isinstance(event, MouseButtonLeftEvent): debug(event.name) #NOTE: copying the list like this before iterating over it, EVERY tick, is highly inefficient, #but currently has to be done because of how new listeners are added to the queue while it is running #(eg when popping cards from a deck). Should be changed. See: http://dr0id.homepage.bluewin.ch/pygame_tutorial08.html #and search for "Watch the iteration" for listener in list(self.listeners): #NOTE: If the weakref has died, it will be #automatically removed, so we don't have #to worry about it. listener.notify(event) #------------------------------------------------------------------------------ class PyGameEventController: """...""" def __init__(self, ev_manager): self.ev_manager = ev_manager self.ev_manager.register_listener(self) self.input_freeze = False #---------------------------------------------------------------------- def notify(self, incoming_event): if isinstance(incoming_event, UserInputFreeze): self.input_freeze = True elif isinstance(incoming_event, UserInputUnFreeze): self.input_freeze = False elif isinstance(incoming_event, TickEvent): #Share some time with other processes, so we don't hog the cpu pygame.time.wait(5) #Handle Pygame Events for event in pygame.event.get(): #If this event manager has an associated PGU GUI app, notify it of the event if self.ev_manager.gui_app: self.ev_manager.gui_app.event(event) #Standard event handling for everything else ev = None if event.type == QUIT: ev = QuitEvent() elif event.type == pygame.MOUSEBUTTONDOWN and not self.input_freeze: if event.button == 1: #Button 1 pos = pygame.mouse.get_pos() ev = MouseButtonLeftEvent(pos) elif event.type == pygame.MOUSEMOTION: pos = pygame.mouse.get_pos() ev = MouseMoveEvent(pos) #Post event to event manager if ev: self.ev_manager.post(ev) #------------------------------------------------------------------------------ class CPUSpinnerController: def __init__(self, ev_manager): self.ev_manager = ev_manager self.ev_manager.register_listener(self) self.clock = pygame.time.Clock() self.cumu_time = 0 self.keep_going = True #---------------------------------------------------------------------- def run(self): if not self.keep_going: raise Exception('dead spinner') while self.keep_going: time_passed = self.clock.tick() fps = self.clock.get_fps() self.cumu_time += time_passed self.ev_manager.post(TickEvent(time_passed, fps)) if self.cumu_time >= 1000: self.cumu_time = 0 self.ev_manager.post(SecondEvent()) pygame.quit() #---------------------------------------------------------------------- def notify(self, event): if isinstance(event, QuitEvent): #this will stop the while loop from running self.keep_going = False rooms.py #Remote imports import pygame #Local imports import config import continents from game_components import * from my_gui import * from pgu import high class Room(object): def __init__(self, screen, ev_manager): self.screen = screen self.ev_manager = ev_manager self.ev_manager.register_listener(self) def notify(self, event): if isinstance(event, TickEvent): pygame.display.set_caption('FPS: ' + str(int(event.fps))) self.render(self.screen) pygame.display.update() def get_highs_table(self): fname = 'high_scores.txt' highs_table = None config.all_highs = high.Highs(fname) if config.game_mode == config.TIME_CHALLENGE: if config.difficulty == config.EASY: highs_table = config.all_highs['time_challenge_easy'] if config.difficulty == config.MED_DIF: highs_table = config.all_highs['time_challenge_med'] if config.difficulty == config.HARD: highs_table = config.all_highs['time_challenge_hard'] if config.difficulty == config.SUPER: highs_table = config.all_highs['time_challenge_super'] elif config.game_mode == config.PLAN_AHEAD: pass return highs_table class TitleScreen(Room): def __init__(self, screen, ev_manager): Room.__init__(self, screen, ev_manager) self.background = pygame.image.load('assets/images/interface/background.jpg').convert() #Initialize #--------------------------------------- self.gui_form = gui.Form() self.gui_app = gui.App(config.gui_theme) self.ev_manager.gui_app = self.gui_app c = gui.Container(align=0,valign=0) #Quit Button #--------------------------------------- b = StartGameButton(ev_manager=self.ev_manager) c.add(b, 0, 0) self.gui_app.init(c) def render(self, surface): surface.blit(self.background, (0, 0)) #GUI self.gui_app.paint(surface) class GameModeRoom(Room): def __init__(self, screen, ev_manager): Room.__init__(self, screen, ev_manager) self.background = pygame.image.load('assets/images/interface/background.jpg').convert() self.create_gui() #Create pgu gui elements def create_gui(self): #Setup #--------------------------------------- self.gui_form = gui.Form() self.gui_app = gui.App(config.gui_theme) self.ev_manager.gui_app = self.gui_app c = gui.Container(align=0,valign=-1) #Mode Relaxed Button #--------------------------------------- b = GameModeRelaxedButton(ev_manager=self.ev_manager) self.b = b print b.rect c.add(b, 0, 200) #Mode Time Challenge Button #--------------------------------------- b = TimeChallengeButton(ev_manager=self.ev_manager) self.b = b print b.rect c.add(b, 0, 250) #Mode Think Ahead Button #--------------------------------------- # b = PlanAheadButton(ev_manager=self.ev_manager) # self.b = b # print b.rect # c.add(b, 0, 300) #Initialize #--------------------------------------- self.gui_app.init(c) def render(self, surface): surface.blit(self.background, (0, 0)) #GUI self.gui_app.paint(surface) class GameRoom(Room): def __init__(self, screen, ev_manager): Room.__init__(self, screen, ev_manager) #Game mode #--------------------------------------- self.new_board_timer = None self.game_mode = config.game_mode config.current_highs = self.get_highs_table() self.highs_dialog = None self.game_over = False #Images #--------------------------------------- self.background = pygame.image.load('assets/images/interface/game screen2-1.jpg').convert() self.logo = pygame.image.load('assets/images/interface/logo_small.png').convert_alpha() self.game_over_text = pygame.image.load('assets/images/interface/text_game_over.png').convert_alpha() self.trip_complete_text = pygame.image.load('assets/images/interface/text_trip_complete.png').convert_alpha() self.zoom_game_over = None self.zoom_trip_complete = None self.fade_out = None #Text #--------------------------------------- self.font = pygame.font.Font(config.font_sans, config.interface_font_size) #Create game components #--------------------------------------- self.continent = self.set_continent(config.continent) self.board = Board(config.board_size, self.ev_manager) self.deck = Deck(self.ev_manager, self.continent) self.map = Map(self.continent) self.longest_trip = 0 #Set pos of game components #--------------------------------------- board_pos = (SCREEN_MARGIN[0], 109) self.board.set_pos(board_pos) map_pos = (config.screen_size[0] - self.map.size[0] - SCREEN_MARGIN[0], 57); self.map.set_pos(map_pos) #Trackers #--------------------------------------- self.game_clock = Chrono(self.ev_manager) self.swap_counter = 0 self.level = 0 #Create gui #--------------------------------------- self.create_gui() #Create initial board #--------------------------------------- self.new_board = self.deck.deal_new_board(self.board) self.ev_manager.post(NewBoardComplete(self.new_board)) def set_continent(self, continent_const): #Set continent based on const if continent_const == config.EUROPE: return continents.Europe() if continent_const == config.AFRICA: return continents.Africa() else: raise Exception('Continent constant not recognized') #Create pgu gui elements def create_gui(self): #Setup #--------------------------------------- self.gui_form = gui.Form() self.gui_app = gui.App(config.gui_theme) self.ev_manager.gui_app = self.gui_app c = gui.Container(align=-1,valign=-1) #Timer Progress bar #--------------------------------------- self.timer_bar = None self.time_increase = None self.minutes_left = None self.seconds_left = None self.timer_text = None if self.game_mode == config.TIME_CHALLENGE: self.time_increase = config.time_challenge_start_time self.timer_bar = gui.ProgressBar(config.time_challenge_start_time,0,config.max_time_bank,width=306) c.add(self.timer_bar, 172, 57) #Connections Progress bar #--------------------------------------- self.connections_bar = None self.connections_bar = gui.ProgressBar(0,0,config.longest_trip_needed,width=306) c.add(self.connections_bar, 172, 83) #Quit Button #--------------------------------------- b = QuitButton(ev_manager=self.ev_manager) c.add(b, 950, 20) #Generate Board Button #--------------------------------------- b = GenerateBoardButton(ev_manager=self.ev_manager, room=self) c.add(b, 500, 20) #Board Size? #--------------------------------------- bs = SetBoardSizeContainer(config.BOARD_LARGE, ev_manager=self.ev_manager, board=self.board) c.add(bs, 640, 20) #Fill Board? #--------------------------------------- t = FillBoardCheckbox(config.fill_board, ev_manager=self.ev_manager) c.add(t, 740, 20) #Darkness? #--------------------------------------- t = UseDarknessCheckbox(config.use_darkness, ev_manager=self.ev_manager) c.add(t, 840, 20) #Initialize #--------------------------------------- self.gui_app.init(c) def advance_level(self): self.level += 1 print 'Advancing to next level' print 'New level: ' + str(self.level) if self.timer_bar: print 'Time increase: ' + str(self.time_increase) self.timer_bar.value += self.time_increase self.time_increase = max(config.min_advance_time, int(self.time_increase * 0.9)) self.board = self.new_board self.new_board = None self.zoom_trip_complete = None self.game_clock.unpause() def notify(self, event): #Tick event if isinstance(event, TickEvent): pygame.display.set_caption('FPS: ' + str(int(event.fps))) self.render(self.screen) pygame.display.update() #Wait to deal new board when advancing levels if self.zoom_trip_complete and self.zoom_trip_complete.finished: self.zoom_trip_complete = None self.ev_manager.post(UnfreezeCards()) self.new_board = self.deck.deal_new_board(self.board) self.ev_manager.post(NewBoardComplete(self.new_board)) #New high score? if self.zoom_game_over and self.zoom_game_over.finished and not self.highs_dialog: if config.current_highs.check(self.level) != None: self.zoom_game_over.visible = False data = 'time:' + str(self.game_clock.time) + ',swaps:' + str(self.swap_counter) self.highs_dialog = HighScoreDialog(score=self.level, data=data, ev_manager=self.ev_manager) self.highs_dialog.open() elif not self.fade_out: self.fade_out = FadeOut(self.ev_manager, config.TITLE_SCREEN) #Second event elif isinstance(event, SecondEvent): if self.timer_bar: if not self.game_clock.paused: self.timer_bar.value -= 1 if self.timer_bar.value <= 0 and not self.game_over: self.ev_manager.post(GameOver()) self.minutes_left = self.timer_bar.value / 60 self.seconds_left = self.timer_bar.value % 60 if self.seconds_left < 10: leading_zero = '0' else: leading_zero = '' self.timer_text = ''.join(['Time Left: ', str(self.minutes_left), ':', leading_zero, str(self.seconds_left)]) #Game over elif isinstance(event, GameOver): self.game_over = True self.zoom_game_over = ZoomImage(self.ev_manager, self.game_over_text) #Trip complete event elif isinstance(event, TripComplete): print 'You did it!' self.game_clock.pause() self.zoom_trip_complete = ZoomImage(self.ev_manager, self.trip_complete_text) self.new_board_timer = Timer(self.ev_manager, 2) self.ev_manager.post(FreezeCards()) print 'Room posted newboardcomplete' #Board Refresh Complete elif isinstance(event, BoardRefreshComplete): if event.board == self.board: print 'Longest trip needed: ' + str(config.longest_trip_needed) print 'Your longest trip: ' + str(self.board.longest_trip) if self.board.longest_trip >= config.longest_trip_needed: self.ev_manager.post(TripComplete()) elif event.board == self.new_board: self.advance_level() self.connections_bar.value = self.board.longest_trip self.connection_text = ' '.join(['Connections:', str(self.board.longest_trip), '/', str(config.longest_trip_needed)]) #CardSwapComplete elif isinstance(event, CardSwapComplete): self.swap_counter += 1 elif isinstance(event, ConfigChangeBoardSize): config.board_size = event.new_size elif isinstance(event, ConfigChangeCardSize): config.card_size = event.new_size elif isinstance(event, ConfigChangeFillBoard): config.fill_board = event.new_value elif isinstance(event, ConfigChangeDarkness): config.use_darkness = event.new_value def render(self, surface): #Background surface.blit(self.background, (0, 0)) #Map self.map.render(surface) #Board self.board.render(surface) #Logo surface.blit(self.logo, (10,10)) #Text connection_text = self.font.render(self.connection_text, True, BLACK) surface.blit(connection_text, (25, 84)) if self.timer_text: timer_text = self.font.render(self.timer_text, True, BLACK) surface.blit(timer_text, (25, 64)) #GUI self.gui_app.paint(surface) if self.zoom_trip_complete: self.zoom_trip_complete.render(surface) if self.zoom_game_over: self.zoom_game_over.render(surface) if self.fade_out: self.fade_out.render(surface) class HighScoresRoom(Room): def __init__(self, screen, ev_manager): Room.__init__(self, screen, ev_manager) self.background = pygame.image.load('assets/images/interface/background.jpg').convert() #Initialize #--------------------------------------- self.gui_app = gui.App(config.gui_theme) self.ev_manager.gui_app = self.gui_app c = gui.Container(align=0,valign=0) #High Scores Table #--------------------------------------- hst = HighScoresTable() c.add(hst, 0, 0) self.gui_app.init(c) def render(self, surface): surface.blit(self.background, (0, 0)) #GUI self.gui_app.paint(surface) game_components.py #Remote imports import pygame from pygame.locals import * import random import operator from copy import copy from math import sqrt, floor #Local imports import config from events import * from matrix import Matrix from textrect import render_textrect, TextRectException from hyphen import hyphenator from textwrap2 import TextWrapper ############################## #CONSTANTS ############################## SCREEN_MARGIN = (10, 10) #Colors BLACK = (0, 0, 0) WHITE = (255, 255, 255) RED = (255, 0, 0) YELLOW = (255, 200, 0) #Directions LEFT = -1 RIGHT = 1 UP = 2 DOWN = -2 #Cards CARD_MARGIN = (10, 10) CARD_PADDING = (2, 2) #Card types BLANK = 0 COUNTRY = 1 TRANSPORT = 2 #Transport types PLANE = 0 TRAIN = 1 CAR = 2 SHIP = 3 class Timer(object): def __init__(self, ev_manager, time_left): self.ev_manager = ev_manager self.ev_manager.register_listener(self) self.time_left = time_left self.paused = False def __repr__(self): return str(self.time_left) def pause(self): self.paused = True def unpause(self): self.paused = False def notify(self, event): #Pause Event if isinstance(event, Pause): self.pause() #Unpause Event elif isinstance(event, Unpause): self.unpause() #Second Event elif isinstance(event, SecondEvent): if not self.paused: self.time_left -= 1 class Chrono(object): def __init__(self, ev_manager, start_time=0): self.ev_manager = ev_manager self.ev_manager.register_listener(self) self.time = start_time self.paused = False def __repr__(self): return str(self.time_left) def pause(self): self.paused = True def unpause(self): self.paused = False def notify(self, event): #Pause Event if isinstance(event, Pause): self.pause() #Unpause Event elif isinstance(event, Unpause): self.unpause() #Second Event elif isinstance(event, SecondEvent): if not self.paused: self.time += 1 class Map(object): def __init__(self, continent): self.map_image = pygame.image.load(continent.map).convert_alpha() self.map_text = pygame.image.load(continent.map_text).convert_alpha() self.pos = (0, 0) self.set_color() self.map_image = pygame.transform.smoothscale(self.map_image, config.map_size) self.size = self.map_image.get_size() def set_pos(self, pos): self.pos = pos def set_color(self): image_pixel_array = pygame.PixelArray(self.map_image) image_pixel_array.replace(config.GRAY1, config.COLOR1) image_pixel_array.replace(config.GRAY2, config.COLOR2) image_pixel_array.replace(config.GRAY3, config.COLOR3) image_pixel_array.replace(config.GRAY4, config.COLOR4) image_pixel_array.replace(config.GRAY5, config.COLOR5)

    Read the article

  • Part 2&ndash;Load Testing In The Cloud

    - by Tarun Arora
    Welcome to Part 2, In Part 1 we discussed the advantages of creating a Test Rig in the cloud, the Azure edge and the Test Rig Topology we want to get to. In Part 2, Let’s start by understanding the components of Azure we’ll be making use of followed by manually putting them together to create the test rig, so… let’s get down dirty start setting up the Test Rig.  What Components of Azure will I be using for building the Test Rig in the Cloud? To run the Test Agents we’ll make use of Windows Azure Compute and to enable communication between Test Controller and Test Agents we’ll make use of Windows Azure Connect.  Azure Connect The Test Controller is on premise and the Test Agents are in the cloud (How will they talk?). To enable communication between the two, we’ll make use of Windows Azure Connect. With Windows Azure Connect, you can use a simple user interface to configure IPsec protected connections between computers or virtual machines (VMs) in your organization’s network, and roles running in Windows Azure. With this you can now join Windows Azure role instances to your domain, so that you can use your existing methods for domain authentication, name resolution, or other domain-wide maintenance actions. For more details refer to an overview of Windows Azure connect. A very useful video explaining everything you wanted to know about Windows Azure connect.  Azure Compute Windows Azure compute provides developers a platform to host and manage applications in Microsoft’s data centres across the globe. A Windows Azure application is built from one or more components called ‘roles.’ Roles come in three different types: Web role, Worker role, and Virtual Machine (VM) role, we’ll be using the Worker role to set up the Test Agents. A very nice blog post discussing the difference between the 3 role types. Developers are free to use the .NET framework or other software that runs on Windows with the Worker role or Web role. Developers can also create applications using languages such as PHP and Java. More on Windows Azure Compute. Each Windows Azure compute instance represents a virtual server... Virtual Machine Size CPU Cores Memory Cost Per Hour Extra Small Shared 768 MB $0.04 Small 1 1.75 GB $0.12 Medium 2 3.50 GB $0.24 Large 4 7.00 GB $0.48 Extra Large 8 14.00 GB $0.96   You might want to review the Windows Azure Pricing FAQ. Let’s Get Started building the Test Rig… Configuration Machine Role Comments VM – 1 Domain Controller for Playpit.com On Premise VM – 2 TFS, Test Controller On Premise VM – 3 Test Agent Cloud   In this blog post I would assume that you have the domain, Team Foundation Server and Test Controller Installed and set up already. If not, please refer to the TFS 2010 Installation Guide and this walkthrough on MSDN to set up your Test Controller. You can also download a preconfigured TFS 2010 VM from Brian Keller's blog, Brian also has some great hands on Labs on TFS 2010 that you may want to explore. I. Lets start building VM – 3: The Test Agent Download the Windows Azure SDK and Tools Open Visual Studio and create a new Windows Azure Project using the Cloud Template                   Choose the Worker Role for reasons explained in the earlier post         The WorkerRole.cs implements the Run() and OnStart() methods, no code changes required. You should be able to compile the project and run it in the compute emulator (The compute emulator should have been installed as part of the Windows Azure Toolkit) on your local machine.                   We will only be making changes to WindowsAzureProject, open ServiceDefinition.csdef. Ensure that the vmsize is small (remember the cost chart above). Import the “Connect” module. I am importing the Connect module because I need to join the Worker role VM to the Playpit domain. <?xml version="1.0" encoding="utf-8"?> <ServiceDefinition name="WindowsAzureProject2" xmlns="http://schemas.microsoft.com/ServiceHosting/2008/10/ServiceDefinition"> <WorkerRole name="WorkerRole1" vmsize="Small"> <Imports> <Import moduleName="Diagnostics" /> <Import moduleName="Connect"/> </Imports> </WorkerRole> </ServiceDefinition> Go to the ServiceConfiguration.Cloud.cscfg and note that settings with key ‘Microsoft.WindowsAzure.Plugins.Connect.%%%%’ have been added to the configuration file. This is because you decided to import the connect module. See the config below. <?xml version="1.0" encoding="utf-8"?> <ServiceConfiguration serviceName="WindowsAzureProject2" xmlns="http://schemas.microsoft.com/ServiceHosting/2008/10/ServiceConfiguration" osFamily="1" osVersion="*"> <Role name="WorkerRole1"> <Instances count="1" /> <ConfigurationSettings> <Setting name="Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString" value="UseDevelopmentStorage=true" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.ActivationToken" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Refresh" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.WaitForConnectivity" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Upgrade" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.EnableDomainJoin" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainFQDN" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainControllerFQDN" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainAccountName" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainPassword" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainOU" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Administrators" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainSiteName" value="" /> </ConfigurationSettings> </Role> </ServiceConfiguration>             Let’s go step by step and understand all the highlighted parameters and where you can find the values for them.       osFamily – By default this is set to 1 (Windows Server 2008 SP2). Change this to 2 if you want the Windows Server 2008 R2 operating system. The Advantage of using osFamily = “2” is that you get Powershell 2.0 rather than Powershell 1.0. In Powershell 2.0 you could simply use “powershell -ExecutionPolicy Unrestricted ./myscript.ps1” and it will work while in Powershell 1.0 you will have to change the registry key by including the following in your command file “reg add HKLM\Software\Microsoft\PowerShell\1\ShellIds\Microsoft.PowerShell /v ExecutionPolicy /d Unrestricted /f” before you can execute any power shell. The other reason you might want to move to os2 is if you wanted IIS 7.5.       Activation Token – To enable communication between the on premise machine and the Windows Azure Worker role VM both need to have the same token. Log on to Windows Azure Management Portal, click on Connect, click on Get Activation Token, this should give you the activation token, copy the activation token to the clipboard and paste it in the configuration file. Note – Later in the blog I’ll be showing you how to install connect on the on premise machine.                       EnableDomainJoin – Set the value to true, ofcourse we want to join the on windows azure worker role VM to the domain.       DomainFQDN, DomainControllerFQDN, DomainAccountName, DomainPassword, DomainOU, Administrators – This information is specific to your domain. I have extracted this information from the ‘service manager’ and ‘Active Directory Users and Computers’. Also, i created a new Domain-OU namely ‘CloudInstances’ so all my cloud instances joined to my domain show up here, this is optional. You can encrypt the DomainPassword – refer to the instructions here. Or hold fire, I’ll be covering that when i come to certificates and encryption in the coming section.       Now once you have filled all this information up, the configuration file should look something like below, <?xml version="1.0" encoding="utf-8"?> <ServiceConfiguration serviceName="WindowsAzureProject2" xmlns="http://schemas.microsoft.com/ServiceHosting/2008/10/ServiceConfiguration" osFamily="2" osVersion="*"> <Role name="WorkerRole1"> <Instances count="1" /> <ConfigurationSettings> <Setting name="Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString" value="UseDevelopmentStorage=true" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.ActivationToken" value="45f55fea-f194-4fbc-b36e-25604faac784" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Refresh" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.WaitForConnectivity" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Upgrade" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.EnableDomainJoin" value="true" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainFQDN" value="play.pit.com" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainControllerFQDN" value="WIN-KUDQMQFGQOL.play.pit.com" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainAccountName" value="playpit\Administrator" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainPassword" value="************************" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainOU" value="OU=CloudInstances, DC=Play, DC=Pit, DC=com" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Administrators" value="Playpit\Administrator" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainSiteName" value="" /> </ConfigurationSettings> </Role> </ServiceConfiguration> Next we will be enabling the Remote Desktop module in to the ServiceDefinition.csdef, we could make changes manually or allow a beautiful wizard to help us make changes. I prefer the second option. So right click on the Windows Azure project and choose Publish       Now once you get the publish wizard, if you haven’t already you would be asked to import your Windows Azure subscription, this is simply the Msdn subscription activation key xml. Once you have done click Next to go to the Settings page and check ‘Enable Remote Desktop for all roles’.       As soon as you do that you get another pop up asking you the details for the user that you would be logging in with (make sure you enter a reasonable expiry date, you do not want the user account to expire today). Notice the more information tag at the bottom, click that to get access to the certificate section. See screen shot below.       From the drop down select the option to create a new certificate        In the pop up window enter the friendly name for your certificate. In my case I entered ‘WAC – Test Rig’ and click ok. This will create a new certificate for you. Click on the view button to see the certificate details. Do you see the Thumbprint, this is the value that will go in the config file (very important). Now click on the Copy to File button to copy the certificate, we will need to import the certificate to the windows Azure Management portal later. So, make sure you save it a safe location.                                Click Finish and enter details of the user you would like to create with permissions for remote desktop access, once you have entered the details on the ‘Remote desktop configuration’ screen click on Ok. From the Publish Windows Azure Wizard screen press Cancel. Cancel because we don’t want to publish the role just yet and Yes because we want to save all the changes in the config file.       Now if you go to the ServiceDefinition.csdef file you will see that the RemoteAccess and RemoteForwarder roles have been imported for you. <?xml version="1.0" encoding="utf-8"?> <ServiceDefinition name="WindowsAzureProject2" xmlns="http://schemas.microsoft.com/ServiceHosting/2008/10/ServiceDefinition"> <WorkerRole name="WorkerRole1" vmsize="Small"> <Imports> <Import moduleName="Diagnostics" /> <Import moduleName="Connect" /> <Import moduleName="RemoteAccess" /> <Import moduleName="RemoteForwarder" /> </Imports> </WorkerRole> </ServiceDefinition> Now go to the ServiceConfiguration.Cloud.cscfg file and you see a whole bunch for setting “Microsoft.WindowsAzure.Plugins.RemoteAccess.%%%” values added for you. <?xml version="1.0" encoding="utf-8"?> <ServiceConfiguration serviceName="WindowsAzureProject2" xmlns="http://schemas.microsoft.com/ServiceHosting/2008/10/ServiceConfiguration" osFamily="2" osVersion="*"> <Role name="WorkerRole1"> <Instances count="1" /> <ConfigurationSettings> <Setting name="Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString" value="UseDevelopmentStorage=true" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.ActivationToken" value="45f55fea-f194-4fbc-b36e-25604faac784" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Refresh" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.WaitForConnectivity" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Upgrade" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.EnableDomainJoin" value="true" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainFQDN" value="play.pit.com" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainControllerFQDN" value="WIN-KUDQMQFGQOL.play.pit.com" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainAccountName" value="playpit\Administrator" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainPassword" value="************************" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainOU" value="OU=CloudInstances, DC=Play, DC=Pit, DC=com" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.Administrators" value="Playpit\Administrator" /> <Setting name="Microsoft.WindowsAzure.Plugins.Connect.DomainSiteName" value="" /> <Setting name="Microsoft.WindowsAzure.Plugins.RemoteAccess.Enabled" value="true" /> <Setting name="Microsoft.WindowsAzure.Plugins.RemoteAccess.AccountUsername" value="Administrator" /> <Setting name="Microsoft.WindowsAzure.Plugins.RemoteAccess.AccountEncryptedPassword" value="MIIBnQYJKoZIhvcNAQcDoIIBjjCCAYoCAQAxggFOMIIBSgIBADAyMB4xHDAaBgNVBAMME1dpbmRvd 3MgQXp1cmUgVG9vbHMCEGa+B46voeO5T305N7TSG9QwDQYJKoZIhvcNAQEBBQAEggEABg4ol5Xol66Ip6QKLbAPWdmD4ae ADZ7aKj6fg4D+ATr0DXBllZHG5Umwf+84Sj2nsPeCyrg3ZDQuxrfhSbdnJwuChKV6ukXdGjX0hlowJu/4dfH4jTJC7sBWS AKaEFU7CxvqYEAL1Hf9VPL5fW6HZVmq1z+qmm4ecGKSTOJ20Fptb463wcXgR8CWGa+1w9xqJ7UmmfGeGeCHQ4QGW0IDSBU6ccg vzF2ug8/FY60K1vrWaCYOhKkxD3YBs8U9X/kOB0yQm2Git0d5tFlIPCBT2AC57bgsAYncXfHvPesI0qs7VZyghk8LVa9g5IqaM Cp6cQ7rmY/dLsKBMkDcdBHuCTAzBgkqhkiG9w0BBwEwFAYIKoZIhvcNAwcECDRVifSXbA43gBApNrp40L1VTVZ1iGag+3O1" /> <Setting name="Microsoft.WindowsAzure.Plugins.RemoteAccess.AccountExpiration" value="2012-11-27T23:59:59.0000000+00:00" /> <Setting name="Microsoft.WindowsAzure.Plugins.RemoteForwarder.Enabled" value="true" /> </ConfigurationSettings> <Certificates> <Certificate name="Microsoft.WindowsAzure.Plugins.RemoteAccess.PasswordEncryption" thumbprint="AA23016CF0BDFC344400B5B82706B608B92E4217" thumbprintAlgorithm="sha1" /> </Certificates> </Role> </ServiceConfiguration>          Okay let’s look at them one at a time,       Enabled - Yes, we would like to enable Remote Access.       AccountUserName – This is the user name you entered while you were on the publish windows azure role screen, as detailed above.       AccountEncrytedPassword – Try and decode that, the certificate is used to encrypt the password you specified for the user account. Remember earlier i said, either use the instructions or wait and i’ll be showing you encryption, now the user account i am using for rdp has the same password as my domain password, so i can simply copy the value of the AccountEncryptedPassword to the DomainPassword as well.       AccountExpiration – This is the expiration as you specified in the wizard earlier, make sure your account does not expire today.       Remote Forwarder – Check out the documentation, below is how I understand it, -- One role in an application that implements a remote desktop connection must import the RemoteForwarder module. The two modules work together to enable the remote desktop connections to role instances. -- If you have multiple roles defined in the service model, it does not matter which role you add the RemoteForwarder module to, but you must add it to only one of the role definitions.       Certificate – Remember the certificate thumbprint from the wizard, the on premise machine and windows azure role machine that need to speak to each other must have the same thumbprint. More on that when we install Windows Azure connect Endpoints on the on premise machine. As i said earlier, in this blog post, I’ll be showing you the manual process so i won’t be scripting any star up tasks to install the test agent or register the test agent with the TFS Server. I’ll be showing you all this cool stuff in the next blog post, that’s because it’s important to understand the manual side of it, it becomes easier for you to troubleshoot in case something fails. Having said that, the changes we have made are sufficient to spin up the Windows Azure Worker Role aka Test Agent VM, have it connected with the play.pit.com domain and have remote access enabled on it. Before we deploy the Test Agent VM we need to set up Windows Azure Connect on the TFS Server. II. Windows Azure Connect: Setting up Connect on VM – 2 i.e. TFS & Test Controller Glad you made it so far, now to enable communication between the on premise TFS/Test Controller and Azure-ed Test Agent we need to enable communication. We have set up the Azure connect module in the Test Agent configuration, now the connect end points need to be enabled on the on premise machines, let’s have a look at how we can do this. Log on to VM – 2 running the TFS Server and Test Controller Log on to the Windows Azure Management Portal and click on Virtual Network Click on Virtual Network, if you already have a subscription you should see the below screen shot, if not, you would be asked to complete the subscription first        Click on Install Local Endpoints from the top left on the panel and you get a url appended with a token id in it, remember the token i showed you earlier, in theory the token you get here should match the token you added to the Test Agent config file.        Copy the url to the clip board and paste it in IE explorer (important, the installation at present only works out of IE and you need to have cookies enabled in order to complete the installation). As stated in the pop up, you can NOT download and run the software later, you need to run it as is, since it contains a token. Once the installation completes you should see the Windows Azure connect icon in the system tray.                         Right click the Azure Connect icon, choose Diagnostics and refer to this link for diagnostic detail terminology. NOTE – Unfortunately I could not see the Windows Azure connect icon in the system tray, a bit of binging with Google revealed that the azure connect icon is only shown when the ‘Windows Azure Connect Endpoint’ Service is started. So go to services.msc and make sure that the service is started, if not start it, unfortunately again, the service did not start for me on a manual start and i realised that one of the dependant services was disabled, you can look at the service dependencies and start them and then start windows azure connect. Bottom line, you need to start Windows Azure connect service before you can proceed. Please refer here on MSDN for more on Troubleshooting Windows Azure connect. (Follow the next step as well)   Now go back to the Windows Azure Management Portal and from Groups and Roles create a new group, lets call it ‘Test Rig’. Make sure you add the VM – 2 (the TFS Server VM where you just installed the endpoint).       Now if you go back to the Azure Connect icon in the system tray and click ‘Refresh Policy’ you will notice that the disconnected status of the icon should change to ready for connection. III. Importing Certificate in to Windows Azure Management Portal But before that you need to import the certificate you created in Step I in to the Windows Azure Management Portal. Log on to the Windows Azure Management Portal and click on ‘Hosted Services, Storage Accounts & CDN’ and then ‘Management Certificates’ followed by Add Certificates as shown in the screen shot below        Browse to the location where you saved the certificate earlier, remember… Refer to Step I in case you forgot.        Now you should be able to see the imported certificate here, make sure the thumbprint of the certificate matches the one you inserted in the config files        IV. Publish Windows Azure Worker Role aka Test Agent Having completed I, II and III, you are ready to publish the Test Agent VM – 3 to the cloud. Go to Visual Studio and right click the Windows Azure project and select Publish. Verify the infomration in the wizard, from the advanced settings tab, you can also enabled capture of intellitrace or profiling information.         Click Next and Click Publish! From the view menu bar select the Windows Azure Activity Log window.       Now you should be able to see the deployment progress in real time.             In the Windows Azure Management Portal, you should also be able to see the progress of creation of a new Worker Role.       Once the deployment is complete you should be able to RDP (go to run prompt type mstsc and in the pop up the machine name) in to the Test Agent Worker Role VM from the Playpit network using the domain admin user account. In case you are unable to log in to the Test Agent using the domain admin user account it means the process of joining the Test Agent to the domain has failed! But the good news is, because you imported the connect module, you can connect to the Test Agent machine using Windows Azure Management Portal and troubleshoot the reason for failure, you will be able to log in with the user name and password you specified in the config file for the keys ‘RemoteAccess.AccountUsername, RemoteAccess.EncryptedPassword (just that enter the password unencrypted)’, fix it or manually join the machine to the domain. Once you have managed to Join the Test Agent VM to the Domain move to the next step.      So, log in to the Test Agent Worker Role VM with the Playpit Domain Administrator and verify that you can log in, the machine is connected to the domain and the connect service is successfully running. If yes, give your self a pat on the back, you are 80% mission accomplished!         Go to the Windows Azure Management Portal and click on Virtual Network, click on Groups and Roles and click on Test Rig, click Edit Group, the edit the Test Rig group you created earlier. In the Connect to section, click on Add to select the worker role you have just deployed. Also, check the ‘Allow connections between endpoints in the group’ with this you will enable to communication between test controller and test agents and test agents/test agents. Click Save.      Now, you are ready to deploy the Test Agent software on the Worker Role Test Agent VM and configure it to work with the Test Controller. V. Configuring VM – 3: Installing Test Agent and Associating Test Agent to Controller Log in to the Worker Role Test Agent VM that you have just successfully deployed, make sure you log in with the domain administrator account. Download the All Agents software from MSDN, ‘en_visual_studio_agents_2010_x86_x64_dvd_509679.iso’, extract the iso and navigate to where you have extracted the iso. In my case, i have extracted the iso to “C:\Resources\Temp\VsAgentSetup”. Open the Test Agent folder and double click on setup.exe. Once you have installed the Test Agent you should reach the configuration window. If you face any issues installing TFS Test Agent on the VM, refer to the walkthrough on MSDN.       Once you have successfully installed the Test Agent software you will need to configure the test agent. Right click the test agent configuration tool and run as a different user. i.e. an Administrator. This is really to run the configuration wizard with elevated privileges (you might have UAC block something's otherwise).        In the run options, you can select ‘service’ you do not need to run the agent as interactive un less you are running coded UI tests. I have specified the domain administrator to connect to the TFS Test Controller. In real life, i would never do that, i would create a separate test user service account for this purpose. But for the blog post, we are using the most powerful user so that any policies or restrictions don’t block you.        Click the Apply Settings button and you should be all green! If not, the summary usually gives helpful error messages that you can resolve and proceed. As per my experience, you may run in to either a permission or a firewall blocking communication issue.        And now the moment of truth! Go to VM –2 open up Visual Studio and from the Test Menu select Manage Test Controller       Mission Accomplished! You should be able to see the Test Agent that you have just configured here,         VI. Creating and Running Load Tests on your brand new Azure-ed Test Rig I have various blog posts on Performance Testing with Visual Studio Ultimate, you can follow the links and videos below, Blog Posts: - Part 1 – Performance Testing using Visual Studio 2010 Ultimate - Part 2 – Performance Testing using Visual Studio 2010 Ultimate - Part 3 – Performance Testing using Visual Studio 2010 Ultimate Videos: - Test Tools Configuration & Settings in Visual Studio - Why & How to Record Web Performance Tests in Visual Studio Ultimate - Goal Driven Load Testing using Visual Studio Ultimate Now that you have created your load tests, there is one last change you need to make before you can run the tests on your Azure Test Rig, create a new Test settings file, and change the Test Execution method to ‘Remote Execution’ and select the test controller you have configured the Worker Role Test Agent against in our case VM – 2 So, go on, fire off a test run and see the results of the test being executed on the Azur-ed Test Rig. Review and What’s next? A quick recap of the benefits of running the Test Rig in the cloud and what i will be covering in the next blog post AND I would love to hear your feedback! Advantages Utilizing the power of Azure compute to run a heavy virtual user load. Benefiting from the Azure flexibility, destroy Test Agents when not in use, takes < 25 minutes to spin up a new Test Agent. Most important test Network Latency, (network latency and speed of connection are two different things – usually network latency is very hard to test), by placing the Test Agents in Microsoft Data centres around the globe, one can actually test the lag in transferring the bytes not because of a slow connection but because the page has been requested from the other side of the globe. Next Steps The process of spinning up the Test Agents in windows Azure is not 100% automated. I am working on the Worker process and power shell scripts to make the role deployment, unattended install of test agent software and registration of the test agent to the test controller automated. In the next blog post I will show you how to make the complete process unattended and automated. Remember to subscribe to http://feeds.feedburner.com/TarunArora. Hope you enjoyed this post, I would love to hear your feedback! If you have any recommendations on things that I should consider or any questions or feedback, feel free to leave a comment. See you in Part III.   Share this post : CodeProject

    Read the article

  • Node.js Adventure - Storage Services and Service Runtime

    - by Shaun
    When I described on how to host a Node.js application on Windows Azure, one of questions might be raised about how to consume the vary Windows Azure services, such as the storage, service bus, access control, etc.. Interact with windows azure services is available in Node.js through the Windows Azure Node.js SDK, which is a module available in NPM. In this post I would like to describe on how to use Windows Azure Storage (a.k.a. WAS) as well as the service runtime.   Consume Windows Azure Storage Let’s firstly have a look on how to consume WAS through Node.js. As we know in the previous post we can host Node.js application on Windows Azure Web Site (a.k.a. WAWS) as well as Windows Azure Cloud Service (a.k.a. WACS). In theory, WAWS is also built on top of WACS worker roles with some more features. Hence in this post I will only demonstrate for hosting in WACS worker role. The Node.js code can be used when consuming WAS when hosted on WAWS. But since there’s no roles in WAWS, the code for consuming service runtime mentioned in the next section cannot be used for WAWS node application. We can use the solution that I created in my last post. Alternatively we can create a new windows azure project in Visual Studio with a worker role, add the “node.exe” and “index.js” and install “express” and “node-sqlserver” modules, make all files as “Copy always”. In order to use windows azure services we need to have Windows Azure Node.js SDK, as knows as a module named “azure” which can be installed through NPM. Once we downloaded and installed, we need to include them in our worker role project and make them as “Copy always”. You can use my “Copy all always” tool mentioned in my last post to update the currently worker role project file. You can also find the source code of this tool here. The source code of Windows Azure SDK for Node.js can be found in its GitHub page. It contains two parts. One is a CLI tool which provides a cross platform command line package for Mac and Linux to manage WAWS and Windows Azure Virtual Machines (a.k.a. WAVM). The other is a library for managing and consuming vary windows azure services includes tables, blobs, queues, service bus and the service runtime. I will not cover all of them but will only demonstrate on how to use tables and service runtime information in this post. You can find the full document of this SDK here. Back to Visual Studio and open the “index.js”, let’s continue our application from the last post, which was working against Windows Azure SQL Database (a.k.a. WASD). The code should looks like this. 1: var express = require("express"); 2: var sql = require("node-sqlserver"); 3:  4: var connectionString = "Driver={SQL Server Native Client 10.0};Server=tcp:ac6271ya9e.database.windows.net,1433;Database=synctile;Uid=shaunxu@ac6271ya9e;Pwd={PASSWORD};Encrypt=yes;Connection Timeout=30;"; 5: var port = 80; 6:  7: var app = express(); 8:  9: app.configure(function () { 10: app.use(express.bodyParser()); 11: }); 12:  13: app.get("/", function (req, res) { 14: sql.open(connectionString, function (err, conn) { 15: if (err) { 16: console.log(err); 17: res.send(500, "Cannot open connection."); 18: } 19: else { 20: conn.queryRaw("SELECT * FROM [Resource]", function (err, results) { 21: if (err) { 22: console.log(err); 23: res.send(500, "Cannot retrieve records."); 24: } 25: else { 26: res.json(results); 27: } 28: }); 29: } 30: }); 31: }); 32:  33: app.get("/text/:key/:culture", function (req, res) { 34: sql.open(connectionString, function (err, conn) { 35: if (err) { 36: console.log(err); 37: res.send(500, "Cannot open connection."); 38: } 39: else { 40: var key = req.params.key; 41: var culture = req.params.culture; 42: var command = "SELECT * FROM [Resource] WHERE [Key] = '" + key + "' AND [Culture] = '" + culture + "'"; 43: conn.queryRaw(command, function (err, results) { 44: if (err) { 45: console.log(err); 46: res.send(500, "Cannot retrieve records."); 47: } 48: else { 49: res.json(results); 50: } 51: }); 52: } 53: }); 54: }); 55:  56: app.get("/sproc/:key/:culture", function (req, res) { 57: sql.open(connectionString, function (err, conn) { 58: if (err) { 59: console.log(err); 60: res.send(500, "Cannot open connection."); 61: } 62: else { 63: var key = req.params.key; 64: var culture = req.params.culture; 65: var command = "EXEC GetItem '" + key + "', '" + culture + "'"; 66: conn.queryRaw(command, function (err, results) { 67: if (err) { 68: console.log(err); 69: res.send(500, "Cannot retrieve records."); 70: } 71: else { 72: res.json(results); 73: } 74: }); 75: } 76: }); 77: }); 78:  79: app.post("/new", function (req, res) { 80: var key = req.body.key; 81: var culture = req.body.culture; 82: var val = req.body.val; 83:  84: sql.open(connectionString, function (err, conn) { 85: if (err) { 86: console.log(err); 87: res.send(500, "Cannot open connection."); 88: } 89: else { 90: var command = "INSERT INTO [Resource] VALUES ('" + key + "', '" + culture + "', N'" + val + "')"; 91: conn.queryRaw(command, function (err, results) { 92: if (err) { 93: console.log(err); 94: res.send(500, "Cannot retrieve records."); 95: } 96: else { 97: res.send(200, "Inserted Successful"); 98: } 99: }); 100: } 101: }); 102: }); 103:  104: app.listen(port); Now let’s create a new function, copy the records from WASD to table service. 1. Delete the table named “resource”. 2. Create a new table named “resource”. These 2 steps ensures that we have an empty table. 3. Load all records from the “resource” table in WASD. 4. For each records loaded from WASD, insert them into the table one by one. 5. Prompt to user when finished. In order to use table service we need the storage account and key, which can be found from the developer portal. Just select the storage account and click the Manage Keys button. Then create two local variants in our Node.js application for the storage account name and key. Since we need to use WAS we need to import the azure module. Also I created another variant stored the table name. In order to work with table service I need to create the storage client for table service. This is very similar as the Windows Azure SDK for .NET. As the code below I created a new variant named “client” and use “createTableService”, specified my storage account name and key. 1: var azure = require("azure"); 2: var storageAccountName = "synctile"; 3: var storageAccountKey = "/cOy9L7xysXOgPYU9FjDvjrRAhaMX/5tnOpcjqloPNDJYucbgTy7MOrAW7CbUg6PjaDdmyl+6pkwUnKETsPVNw=="; 4: var tableName = "resource"; 5: var client = azure.createTableService(storageAccountName, storageAccountKey); Now create a new function for URL “/was/init” so that we can trigger it through browser. Then in this function we will firstly load all records from WASD. 1: app.get("/was/init", function (req, res) { 2: // load all records from windows azure sql database 3: sql.open(connectionString, function (err, conn) { 4: if (err) { 5: console.log(err); 6: res.send(500, "Cannot open connection."); 7: } 8: else { 9: conn.queryRaw("SELECT * FROM [Resource]", function (err, results) { 10: if (err) { 11: console.log(err); 12: res.send(500, "Cannot retrieve records."); 13: } 14: else { 15: if (results.rows.length > 0) { 16: // begin to transform the records into table service 17: } 18: } 19: }); 20: } 21: }); 22: }); When we succeed loaded all records we can start to transform them into table service. First I need to recreate the table in table service. This can be done by deleting and creating the table through table client I had just created previously. 1: app.get("/was/init", function (req, res) { 2: // load all records from windows azure sql database 3: sql.open(connectionString, function (err, conn) { 4: if (err) { 5: console.log(err); 6: res.send(500, "Cannot open connection."); 7: } 8: else { 9: conn.queryRaw("SELECT * FROM [Resource]", function (err, results) { 10: if (err) { 11: console.log(err); 12: res.send(500, "Cannot retrieve records."); 13: } 14: else { 15: if (results.rows.length > 0) { 16: // begin to transform the records into table service 17: // recreate the table named 'resource' 18: client.deleteTable(tableName, function (error) { 19: client.createTableIfNotExists(tableName, function (error) { 20: if (error) { 21: error["target"] = "createTableIfNotExists"; 22: res.send(500, error); 23: } 24: else { 25: // transform the records 26: } 27: }); 28: }); 29: } 30: } 31: }); 32: } 33: }); 34: }); As you can see, the azure SDK provide its methods in callback pattern. In fact, almost all modules in Node.js use the callback pattern. For example, when I deleted a table I invoked “deleteTable” method, provided the name of the table and a callback function which will be performed when the table had been deleted or failed. Underlying, the azure module will perform the table deletion operation in POSIX async threads pool asynchronously. And once it’s done the callback function will be performed. This is the reason we need to nest the table creation code inside the deletion function. If we perform the table creation code after the deletion code then they will be invoked in parallel. Next, for each records in WASD I created an entity and then insert into the table service. Finally I send the response to the browser. Can you find a bug in the code below? I will describe it later in this post. 1: app.get("/was/init", function (req, res) { 2: // load all records from windows azure sql database 3: sql.open(connectionString, function (err, conn) { 4: if (err) { 5: console.log(err); 6: res.send(500, "Cannot open connection."); 7: } 8: else { 9: conn.queryRaw("SELECT * FROM [Resource]", function (err, results) { 10: if (err) { 11: console.log(err); 12: res.send(500, "Cannot retrieve records."); 13: } 14: else { 15: if (results.rows.length > 0) { 16: // begin to transform the records into table service 17: // recreate the table named 'resource' 18: client.deleteTable(tableName, function (error) { 19: client.createTableIfNotExists(tableName, function (error) { 20: if (error) { 21: error["target"] = "createTableIfNotExists"; 22: res.send(500, error); 23: } 24: else { 25: // transform the records 26: for (var i = 0; i < results.rows.length; i++) { 27: var entity = { 28: "PartitionKey": results.rows[i][1], 29: "RowKey": results.rows[i][0], 30: "Value": results.rows[i][2] 31: }; 32: client.insertEntity(tableName, entity, function (error) { 33: if (error) { 34: error["target"] = "insertEntity"; 35: res.send(500, error); 36: } 37: else { 38: console.log("entity inserted"); 39: } 40: }); 41: } 42: // send the 43: console.log("all done"); 44: res.send(200, "All done!"); 45: } 46: }); 47: }); 48: } 49: } 50: }); 51: } 52: }); 53: }); Now we can publish it to the cloud and have a try. But normally we’d better test it at the local emulator first. In Node.js SDK there are three build-in properties which provides the account name, key and host address for local storage emulator. We can use them to initialize our table service client. We also need to change the SQL connection string to let it use my local database. The code will be changed as below. 1: // windows azure sql database 2: //var connectionString = "Driver={SQL Server Native Client 10.0};Server=tcp:ac6271ya9e.database.windows.net,1433;Database=synctile;Uid=shaunxu@ac6271ya9e;Pwd=eszqu94XZY;Encrypt=yes;Connection Timeout=30;"; 3: // sql server 4: var connectionString = "Driver={SQL Server Native Client 11.0};Server={.};Database={Caspar};Trusted_Connection={Yes};"; 5:  6: var azure = require("azure"); 7: var storageAccountName = "synctile"; 8: var storageAccountKey = "/cOy9L7xysXOgPYU9FjDvjrRAhaMX/5tnOpcjqloPNDJYucbgTy7MOrAW7CbUg6PjaDdmyl+6pkwUnKETsPVNw=="; 9: var tableName = "resource"; 10: // windows azure storage 11: //var client = azure.createTableService(storageAccountName, storageAccountKey); 12: // local storage emulator 13: var client = azure.createTableService(azure.ServiceClient.DEVSTORE_STORAGE_ACCOUNT, azure.ServiceClient.DEVSTORE_STORAGE_ACCESS_KEY, azure.ServiceClient.DEVSTORE_TABLE_HOST); Now let’s run the application and navigate to “localhost:12345/was/init” as I hosted it on port 12345. We can find it transformed the data from my local database to local table service. Everything looks fine. But there is a bug in my code. If we have a look on the Node.js command window we will find that it sent response before all records had been inserted, which is not what I expected. The reason is that, as I mentioned before, Node.js perform all IO operations in non-blocking model. When we inserted the records we executed the table service insert method in parallel, and the operation of sending response was also executed in parallel, even though I wrote it at the end of my logic. The correct logic should be, when all entities had been copied to table service with no error, then I will send response to the browser, otherwise I should send error message to the browser. To do so I need to import another module named “async”, which helps us to coordinate our asynchronous code. Install the module and import it at the beginning of the code. Then we can use its “forEach” method for the asynchronous code of inserting table entities. The first argument of “forEach” is the array that will be performed. The second argument is the operation for each items in the array. And the third argument will be invoked then all items had been performed or any errors occurred. Here we can send our response to browser. 1: app.get("/was/init", function (req, res) { 2: // load all records from windows azure sql database 3: sql.open(connectionString, function (err, conn) { 4: if (err) { 5: console.log(err); 6: res.send(500, "Cannot open connection."); 7: } 8: else { 9: conn.queryRaw("SELECT * FROM [Resource]", function (err, results) { 10: if (err) { 11: console.log(err); 12: res.send(500, "Cannot retrieve records."); 13: } 14: else { 15: if (results.rows.length > 0) { 16: // begin to transform the records into table service 17: // recreate the table named 'resource' 18: client.deleteTable(tableName, function (error) { 19: client.createTableIfNotExists(tableName, function (error) { 20: if (error) { 21: error["target"] = "createTableIfNotExists"; 22: res.send(500, error); 23: } 24: else { 25: async.forEach(results.rows, 26: // transform the records 27: function (row, callback) { 28: var entity = { 29: "PartitionKey": row[1], 30: "RowKey": row[0], 31: "Value": row[2] 32: }; 33: client.insertEntity(tableName, entity, function (error) { 34: if (error) { 35: callback(error); 36: } 37: else { 38: console.log("entity inserted."); 39: callback(null); 40: } 41: }); 42: }, 43: // send reponse 44: function (error) { 45: if (error) { 46: error["target"] = "insertEntity"; 47: res.send(500, error); 48: } 49: else { 50: console.log("all done"); 51: res.send(200, "All done!"); 52: } 53: } 54: ); 55: } 56: }); 57: }); 58: } 59: } 60: }); 61: } 62: }); 63: }); Run it locally and now we can find the response was sent after all entities had been inserted. Query entities against table service is simple as well. Just use the “queryEntity” method from the table service client and providing the partition key and row key. We can also provide a complex query criteria as well, for example the code here. In the code below I queried an entity by the partition key and row key, and return the proper localization value in response. 1: app.get("/was/:key/:culture", function (req, res) { 2: var key = req.params.key; 3: var culture = req.params.culture; 4: client.queryEntity(tableName, culture, key, function (error, entity) { 5: if (error) { 6: res.send(500, error); 7: } 8: else { 9: res.json(entity); 10: } 11: }); 12: }); And then tested it on local emulator. Finally if we want to publish this application to the cloud we should change the database connection string and storage account. For more information about how to consume blob and queue service, as well as the service bus please refer to the MSDN page.   Consume Service Runtime As I mentioned above, before we published our application to the cloud we need to change the connection string and account information in our code. But if you had played with WACS you should have known that the service runtime provides the ability to retrieve configuration settings, endpoints and local resource information at runtime. Which means we can have these values defined in CSCFG and CSDEF files and then the runtime should be able to retrieve the proper values. For example we can add some role settings though the property window of the role, specify the connection string and storage account for cloud and local. And the can also use the endpoint which defined in role environment to our Node.js application. In Node.js SDK we can get an object from “azure.RoleEnvironment”, which provides the functionalities to retrieve the configuration settings and endpoints, etc.. In the code below I defined the connection string variants and then use the SDK to retrieve and initialize the table client. 1: var connectionString = ""; 2: var storageAccountName = ""; 3: var storageAccountKey = ""; 4: var tableName = ""; 5: var client; 6:  7: azure.RoleEnvironment.getConfigurationSettings(function (error, settings) { 8: if (error) { 9: console.log("ERROR: getConfigurationSettings"); 10: console.log(JSON.stringify(error)); 11: } 12: else { 13: console.log(JSON.stringify(settings)); 14: connectionString = settings["SqlConnectionString"]; 15: storageAccountName = settings["StorageAccountName"]; 16: storageAccountKey = settings["StorageAccountKey"]; 17: tableName = settings["TableName"]; 18:  19: console.log("connectionString = %s", connectionString); 20: console.log("storageAccountName = %s", storageAccountName); 21: console.log("storageAccountKey = %s", storageAccountKey); 22: console.log("tableName = %s", tableName); 23:  24: client = azure.createTableService(storageAccountName, storageAccountKey); 25: } 26: }); In this way we don’t need to amend the code for the configurations between local and cloud environment since the service runtime will take care of it. At the end of the code we will listen the application on the port retrieved from SDK as well. 1: azure.RoleEnvironment.getCurrentRoleInstance(function (error, instance) { 2: if (error) { 3: console.log("ERROR: getCurrentRoleInstance"); 4: console.log(JSON.stringify(error)); 5: } 6: else { 7: console.log(JSON.stringify(instance)); 8: if (instance["endpoints"] && instance["endpoints"]["nodejs"]) { 9: var endpoint = instance["endpoints"]["nodejs"]; 10: app.listen(endpoint["port"]); 11: } 12: else { 13: app.listen(8080); 14: } 15: } 16: }); But if we tested the application right now we will find that it cannot retrieve any values from service runtime. This is because by default, the entry point of this role was defined to the worker role class. In windows azure environment the service runtime will open a named pipeline to the entry point instance, so that it can connect to the runtime and retrieve values. But in this case, since the entry point was worker role and the Node.js was opened inside the role, the named pipeline was established between our worker role class and service runtime, so our Node.js application cannot use it. To fix this problem we need to open the CSDEF file under the azure project, add a new element named Runtime. Then add an element named EntryPoint which specify the Node.js command line. So that the Node.js application will have the connection to service runtime, then it’s able to read the configurations. Start the Node.js at local emulator we can find it retrieved the connections, storage account for local. And if we publish our application to azure then it works with WASD and storage service through the configurations for cloud.   Summary In this post I demonstrated how to use Windows Azure SDK for Node.js to interact with storage service, especially the table service. I also demonstrated on how to use WACS service runtime, how to retrieve the configuration settings and the endpoint information. And in order to make the service runtime available to my Node.js application I need to create an entry point element in CSDEF file and set “node.exe” as the entry point. I used five posts to introduce and demonstrate on how to run a Node.js application on Windows platform, how to use Windows Azure Web Site and Windows Azure Cloud Service worker role to host our Node.js application. I also described how to work with other services provided by Windows Azure platform through Windows Azure SDK for Node.js. Node.js is a very new and young network application platform. But since it’s very simple and easy to learn and deploy, as well as, it utilizes single thread non-blocking IO model, Node.js became more and more popular on web application and web service development especially for those IO sensitive projects. And as Node.js is very good at scaling-out, it’s more useful on cloud computing platform. Use Node.js on Windows platform is new, too. The modules for SQL database and Windows Azure SDK are still under development and enhancement. It doesn’t support SQL parameter in “node-sqlserver”. It does support using storage connection string to create the storage client in “azure”. But Microsoft is working on make them easier to use, working on add more features and functionalities.   PS, you can download the source code here. You can download the source code of my “Copy all always” tool here.   Hope this helps, Shaun All documents and related graphics, codes are provided "AS IS" without warranty of any kind. Copyright © Shaun Ziyan Xu. This work is licensed under the Creative Commons License.

    Read the article

  • Validating the SharePoint InputFormTextBox / RichText Editor using JavaScript

    - by Jignesh Gangajaliya
    In the previous post I mentioned about manipulating SharePoint PeoplePicker control using JavaScript, in this post I will explain how to validate the InputFormTextBox contol using JavaScript. Here is the nice post by Becky Isserman on why not to use RequiredFieldValdator or InputFormRequiredFieldValidator with InputFormTextbox. function ValidateComments() {     //retrieve the text from rich text editor.     var text = RTE_GetRichEditTextOnly("<%= rteComments.ClientID %>");     if (text != "")     {         return true;     }     else     {         alert('Please enter your comments.');         //set focus back to the rich text editor.         RTE_GiveEditorFocus("<%= rteComments.ClientID %>");         return false;     }     return true; } <SharePoint:InputFormTextBox ID="rteComments" runat="server" RichText="true" RichTextMode="Compatible" Rows="10" TextMode="MultiLine" CausesValidation="true" ></SharePoint:InputFormTextBox> <asp:Button ID="btnSubmit" runat="server" Text="Submit" OnClick="btnSubmit_Click" OnClientClick="return ValidateComments()" CausesValidation="true" /> - Jignesh

    Read the article

  • How to Fix “Error occurred in deployment step ‘Activate Features’: System.TimeoutException:”

    - by ybbest
    Problem: When deploying a SharePoint2013 workflow using Visual Studio, I got the following Error: Error occurred in deployment step ‘Activate Features’: System.TimeoutException: The HTTP request has timed out after 20000 milliseconds. —> System.Net.WebException: The request was aborted: The request was canceled. at System.Net.HttpWebRequest.EndGetResponse(IAsyncResult asyncResult) at Microsoft.Workflow.Client.HttpGetResponseAsyncResult`1.OnGotResponse(IAsyncResult result) — End of inner exception stack trace — at Microsoft.Workflow.Common.AsyncResult.End[TAsyncResult](IAsyncResult result) at Microsoft.Workflow.Client.Ht Analysis: After reading AC’s blogpost and I find out the issue is to do with the service bus. Then I found out the following services are not started Solution: So I start the Service Bus Gateway and Service Bus Message Broker and the problem goes away. References: SharePoint 2013 Workflow – Advanced Workflow Debugging with Fiddler

    Read the article

  • SQL SERVER – Spatial Database Queries – What About BLOB – T-SQL Tuesday #006

    - by pinaldave
    Michael Coles is one of the most interesting book authors I have ever met. He has a flair of writing complex stuff in a simple language. There are a very few people like that.  I really enjoyed reading his recent book, Expert SQL Server 2008 Encryption. I strongly suggest taking a look at it. This blog is written in response to T-SQL Tuesday #006: “What About BLOB? by Michael Coles. Spatial Database is my favorite subject. Since I did my TechEd India 2010 presentation, I have enjoyed this subject a lot. Before I continue this blog post, there are a few other blog posts, so I suggest you read them.  To help build the environment run the queries, I am going to present them in this single blog post. SQL SERVER – What is Spatial Database? – Developing with SQL Server Spatial and Deep Dive into Spatial Indexing This blog post explains the basics of Spatial Database and also provides a good introduction to Indexing concept. SQL SERVER – World Shapefile Download and Upload to Database – Spatial Database This blog post will enable you with how to load the shape file into database. SQL SERVER – Spatial Database Definition and Research Documents This blog post links to the white paper about Spatial Database written by Microsoft experts. SQL SERVER – Introduction to Spatial Coordinate Systems: Flat Maps for a Round Planet This blog post links to the white paper explaining coordinate system, as written by Microsoft experts. After reading the above listed blog posts, I am very confident that you are ready to run the following script. Once you create a database using the World Shapefile, as mentioned in the second link above,you can display the image of India just like the following. Please note that this is not an accurate political map. The boundary of this map has many errors and it is just a representation. You can run the following query to generate the map of India from the database spatial which you have created after following the instructions here. USE Spatial GO -- India Map SELECT [CountryName] ,[BorderAsGeometry] ,[Border] FROM [Spatial].[dbo].[Countries] WHERE Countryname = 'India' GO Now, let us find the longitude and latitude of the two major IT cities of India, Hyderabad and Bangalore. I find their values as the following: the values of longitude-latitude for Bangalore is 77.5833300000 13.0000000000; for Hyderabad, longitude-latitude is 78.4675900000 17.4531200000. Now, let us try to put these values on the India Map and see their location. -- Bangalore DECLARE @GeoLocation GEOGRAPHY SET @GeoLocation = GEOGRAPHY::STPointFromText('POINT(77.5833300000 13.0000000000)',4326).STBuffer(20000); -- Hyderabad DECLARE @GeoLocation1 GEOGRAPHY SET @GeoLocation1 = GEOGRAPHY::STPointFromText('POINT(78.4675900000 17.4531200000)',4326).STBuffer(20000); -- Bangalore and Hyderabad on Map of India SELECT name, [GeoLocation] FROM [IndiaGeoNames] I WHERE I.[GeoLocation].STDistance(@GeoLocation) <= 0 UNION ALL SELECT name, [GeoLocation] FROM [IndiaGeoNames] I WHERE I.[GeoLocation].STDistance(@GeoLocation1) <= 0 UNION ALL SELECT '',[Border] FROM [Spatial].[dbo].[Countries] WHERE Countryname = 'India' GO Now let us quickly draw a straight line between them. DECLARE @GeoLocation GEOGRAPHY SET @GeoLocation = GEOGRAPHY::STPointFromText('POINT(78.4675900000 17.4531200000)',4326).STBuffer(10000); DECLARE @GeoLocation1 GEOGRAPHY SET @GeoLocation1 = GEOGRAPHY::STPointFromText('POINT(77.5833300000 13.0000000000)',4326).STBuffer(10000); DECLARE @GeoLocation2 GEOGRAPHY SET @GeoLocation2 = GEOGRAPHY::STGeomFromText('LINESTRING(78.4675900000 17.4531200000, 77.5833300000 13.0000000000)',4326) SELECT name, [GeoLocation] FROM [IndiaGeoNames] I WHERE I.[GeoLocation].STDistance(@GeoLocation) <= 0 UNION ALL SELECT name, [GeoLocation] FROM [IndiaGeoNames] I1 WHERE I1.[GeoLocation].STDistance(@GeoLocation1) <= 0 UNION ALL SELECT '' name, @GeoLocation2 UNION ALL SELECT '',[Border] FROM [Spatial].[dbo].[Countries] WHERE Countryname = 'India' GO Let us use the distance function of the spatial database and find the straight line distance between this two cities. -- Distance Between Hyderabad and Bangalore DECLARE @GeoLocation GEOGRAPHY SET @GeoLocation = GEOGRAPHY::STPointFromText('POINT(78.4675900000 17.4531200000)',4326) DECLARE @GeoLocation1 GEOGRAPHY SET @GeoLocation1 = GEOGRAPHY::STPointFromText('POINT(77.5833300000 13.0000000000)',4326) SELECT @GeoLocation.STDistance(@GeoLocation1)/1000 'KM'; GO The result of above query is as displayed in following image. As per SQL Server, the distance between these two cities is 501 KM, but according to what I know, the distance between those two cities is around 562 KM by road. However, please note that roads are not straight and they have lots of turns, whereas this is a straight-line distance. What would be more accurate is the distance between these two cities by air travel. When we look at the air travel distance between Bangalore and Hyderabad, the total distance covered is 495 KM, which is very close to what SQL Server has estimated, which is 501 KM. Bravo! SQL Server has accurately provided the distance between two of the cities. SQL Server Spatial Database can be very useful simply because it is very easy to use, as demonstrated above. I appreciate your comments, so let me know what your thoughts and opinions about this are. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: Pinal Dave, SQL, SQL Authority, SQL Query, SQL Scripts, SQL Server, SQL Tips and Tricks, T SQL, Technology Tagged: Spatial Database

    Read the article

  • SQL SERVER – Interview Questions and Answers – Frequently Asked Questions – Complete Downloadable List – Day 0 of 31

    - by pinaldave
    This blog post is running list of the blog posts in the series of Interview Questions and Answers. At the end of the 31st day of the month, a FREE PDF will be posted here which can be downloadable for offline review. Please scroll below to see latest post for the day. SQL SERVER – Interview Questions and Answers – Frequently Asked Questions – Introduction – Day 1 of 31 In this very first blog post – various aspect of the interview questions and answers are discussed. Some people like the subject for their helpful hints and thought provoking subject, and others dislike these posts because they feel it is nothing more than cheating.  I’d like to discuss the pros and cons of a Question and Answer format here. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: Pinal Dave, PostADay, SQL, SQL Authority, SQL Documentation, SQL Download, SQL Interview Questions and Answers, SQL Query, SQL Server, SQL Tips and Tricks, T SQL, Technology

    Read the article

  • Integrating Twitter Into An ASP.NET Website Using OAuth

    Earlier this year I wrote an article about <a href="http://www.twitterizer.net/">Twitterizer</a>, an open-source .NET library that can be used to integrate your application with <a href="http://twitter.com/">Twitter</a>. Using Twitterizer you can allow your visitors to post tweets, view their timeline, and much more, all without leaving your website. The original article, <a href="http://www.4guysfromrolla.com/articles/021710-1.aspx">Integrating Twitter Into An ASP.NET Website</a>, showed how to post tweets and view a timeline to a particular Twitter account using Twitterizer 1.0. To post a tweet to a specific account, Twitterizer 1.0 uses <i>basic authentication</i>. Basic authentication is a very simple

    Read the article

  • ASP.NET Chart Control - During a PostBack

    - by Guilherme Cardoso
    To use the Chart control from a PostBack is necessary to modify the ChartImg.axd HttpHandler, otherwise we'll get the error message: Error executing child request for ChartImg.axd In Web.Config search the line: <add path = "ChartImg.axd" verb = "GET,HEAD" type = "System.Web.UI.DataVisualization.Charting.ChartHttpHandler, System.Web.DataVisualization, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" <Add path = "ChartImg.axd" verb = "GET, HEAD" type = "System.Web.UI.DataVisualization.Charting.ChartHttpHandler, System.Web.DataVisualization, Version = 3.5.0.0, Culture = neutral, PublicKeyToken = 31bf3856ad364e35 " validate = "false" /> Validate = "false" />   Change to: <add path = "ChartImg.axd" verb = "GET,HEAD,POST" type = "System.Web.UI.DataVisualization.Charting.ChartHttpHandler, System.Web.DataVisualization, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" <Add path = "ChartImg.axd" verb = "GET, HEAD, POST" type = "System.Web.UI.DataVisualization.Charting.ChartHttpHandler, System.Web.DataVisualization, Version = 3.5.0.0, Culture = Neutral, PublicKeyToken = 31bf3856ad364e35 " validate = "false" /> validate = "false" /> The attribute that we are adding is the Post.  For those not familiar with this control is very useful for creating graphics. You can see more information here .

    Read the article

  • T-SQL Tuesday #006: LOB, row-overflow and locking behavior

    - by Michael Zilberstein
    This post is my contribution to T-SQL Tuesday #006 , hosted this time by Michael Coles . Actually this post was born last Thursday when I attended Kalen Delaney's "Deep dive into SQL Server Internals" seminar in Tel-Aviv. I asked question, Kalen didn't have answer at hand, so during a break I created demo in order to check certain behavior. Demo goes later in this post but first small teaser. I have MyTable table with 10 rows. I take 2 rows that reside on different pages. In first session...(read more)

    Read the article

  • Accessing Server-Side Data from Client Script: Using WCF Services with jQuery and the ASP.NET Ajax Library

    Today's websites commonly exchange information between the browser and the web server using Ajax techniques - the browser executes JavaScript code typically in response to the page loading or some user action. This JavaScript makes an asynchronous HTTP request to the server. which then processes the request and, perhaps, returns data that the browser can then seamlessly integrate into the web page. Two earlier articles - Accessing JSON Data From an ASP.NET Page Using jQuery and Using Ajax Web Services, Script References, and jQuery, looked at using both jQuery and the ASP.NET Ajax Library on the browser to initiate an Ajax request and both ASP.NET pages and Ajax Web Services as the entities on the web server responsible for servicing such Ajax requests. This article continues our examination of techniques for implementing lightweight Ajax scenarios in an ASP.NET website. Specifically, it examines how to use the Windows Communication Foundation, or WCF, to serve data from the web server and how to use both the ASP.NET Ajax Library and jQuery to consume such services from the client-side. Read on to learn more! Read More >

    Read the article

  • ASP.NET Web API - Screencast series Part 3: Delete and Update

    - by Jon Galloway
    We're continuing a six part series on ASP.NET Web API that accompanies the getting started screencast series. This is an introductory screencast series that walks through from File / New Project to some more advanced scenarios like Custom Validation and Authorization. The screencast videos are all short (3-5 minutes) and the sample code for the series is both available for download and browsable online. I did the screencasts, but the samples were written by the ASP.NET Web API team. In Part 1 we looked at what ASP.NET Web API is, why you'd care, did the File / New Project thing, and did some basic HTTP testing using browser F12 developer tools. In Part 2 we started to build up a sample that returns data from a repository in JSON format via GET methods. In Part 3, we'll start to modify data on the server using DELETE and POST methods. So far we've been looking at GET requests, and the difference between standard browsing in a web browser and navigating an HTTP API isn't quite as clear. Delete is where the difference becomes more obvious. With a "traditional" web page, to delete something'd probably have a form that POSTs a request back to a controller that needs to know that it's really supposed to be deleting something even though POST was really designed to create things, so it does the work and then returns some HTML back to the client that says whether or not the delete succeeded. There's a good amount of plumbing involved in communicating between client and server. That gets a lot easier when we just work with the standard HTTP DELETE verb. Here's how the server side code works: public Comment DeleteComment(int id) { Comment comment; if (!repository.TryGet(id, out comment)) throw new HttpResponseException(HttpStatusCode.NotFound); repository.Delete(id); return comment; } If you look back at the GET /api/comments code in Part 2, you'll see that they start the exact same because the use cases are kind of similar - we're looking up an item by id and either displaying it or deleting it. So the only difference is that this method deletes the comment once it finds it. We don't need to do anything special to handle cases where the id isn't found, as the same HTTP 404 handling works fine here, too. Pretty much all "traditional" browsing uses just two HTTP verbs: GET and POST, so you might not be all that used to DELETE requests and think they're hard. Not so! Here's the jQuery method that calls the /api/comments with the DELETE verb: $(function() { $("a.delete").live('click', function () { var id = $(this).data('comment-id'); $.ajax({ url: "/api/comments/" + id, type: 'DELETE', cache: false, statusCode: { 200: function(data) { viewModel.comments.remove( function(comment) { return comment.ID == data.ID; } ); } } }); return false; }); }); So in order to use the DELETE verb instead of GET, we're just using $.ajax() and setting the type to DELETE. Not hard. But what's that statusCode business? Well, an HTTP status code of 200 is an OK response. Unless our Web API method sets another status (such as by throwing the Not Found exception we saw earlier), the default response status code is HTTP 200 - OK. That makes the jQuery code pretty simple - it calls the Delete action, and if it gets back an HTTP 200, the server-side delete was successful so the comment can be deleted. Adding a new comment uses the POST verb. It starts out looking like an MVC controller action, using model binding to get the new comment from JSON data into a c# model object to add to repository, but there are some interesting differences. public HttpResponseMessage<Comment> PostComment(Comment comment) { comment = repository.Add(comment); var response = new HttpResponseMessage<Comment>(comment, HttpStatusCode.Created); response.Headers.Location = new Uri(Request.RequestUri, "/api/comments/" + comment.ID.ToString()); return response; } First off, the POST method is returning an HttpResponseMessage<Comment>. In the GET methods earlier, we were just returning a JSON payload with an HTTP 200 OK, so we could just return the  model object and Web API would wrap it up in an HttpResponseMessage with that HTTP 200 for us (much as ASP.NET MVC controller actions can return strings, and they'll be automatically wrapped in a ContentResult). When we're creating a new comment, though, we want to follow standard REST practices and return the URL that points to the newly created comment in the Location header, and we can do that by explicitly creating that HttpResposeMessage and then setting the header information. And here's a key point - by using HTTP standard status codes and headers, our response payload doesn't need to explain any context - the client can see from the status code that the POST succeeded, the location header tells it where to get it, and all it needs in the JSON payload is the actual content. Note: This is a simplified sample. Among other things, you'll need to consider security and authorization in your Web API's, and especially in methods that allow creating or deleting data. We'll look at authorization in Part 6. As for security, you'll want to consider things like mass assignment if binding directly to model objects, etc. In Part 4, we'll extend on our simple querying methods form Part 2, adding in support for paging and querying.

    Read the article

  • Using stored procedures with Entity Framework in an ASP.Net application

    - by nikolaosk
    This is going to be the third post of a series of posts regarding ASP.Net and the Entity Framework and how we can use Entity Framework to access our datastore. You can find the first one here and the second one here . I have a post regarding ASP.Net and EntityDataSource. You can read it here .I have 3 more posts on Profiling Entity Framework applications. You can have a look at them here , here and here . In this post I will show you how to select,insert,update,delete data in the database using EF...(read more)

    Read the article

  • SyFy Channel Original Movie Title Generator

    - by Most Valuable Yak (Rob Volk)
    Saw this linked on reddit today and couldn't resist going through all the combinations: create table #pre(name varchar(20))create table #post(name varchar(20), pre varchar(10))insert #pre select 'Dino' union all select'Alien' union all select'Shark' union all select'Raptor' union all select'Tractor' union all select'Arachno' union all select'Cyber' union all select'Robo' union all select'Choco' union all select'Chupa' union all select'Grizzly' union all select'Mega' union all select'Were' union all select'Sabre' union all select'Man' insert #post select 'dactyl','a' union all select'pus','to' union all select'conda','a' union all select'droid',null union all select'dile','o' union all select'bear',null union all select'vampire',null union all select'squito',null union all select'saurus','a' union all select'wolf',null union all select'ghost',null union all select'viper',null union all select'cabra','a' union all select'yeti',null union all select'shark',null select a.name +case when right(a.name,1) not like '[aeiouy]' and b.pre is not null then b.pre else '' end +b.namefrom #pre a cross join #post bwhere a.name<>b.name -- optional, to eliminate the "SharkShark" optionorder by 1  Which one is your favorite?  I like most of the -squito versions, especially Chupasquito and Grizzlysquito.

    Read the article

  • WebGL, security, and Microsoft

    - by 3412132
    I was writing a post about a link I saw, but realized it was also about what companies do to this industry, so I'd like to ask your opinions on that first (the original post is below). Is it ok for companies to act childish (not wanting to share, not-invented-here syndrome, etc)? ORIGINAL POST: http://news.cnet.com/8301-30685_3-20071726-264/microsoft-declares-webgl-harmful-to-security/ What gives? I understand they're making some real points here, but haven't they been doing similar things with ActiveX? Also who are they to talk when their browser has more security problems than modern browsers do?

    Read the article

  • Concurrent Business Events

    - by Manoj Madhusoodanan
    This blog describes the various business events related to concurrent requests.In the concurrent program definition screen we can see the various business events which are attached to concurrent processing. Following are the actual definition of above business events. Each event will have following parameters. Create subscriptions to above business events.Before testing enable profile option 'Concurrent: Business Intelligence Integration Enable' to Yes. ExampleI have created a scenario.Whenever my concurrent request completes normally I want to send out file as attachment to my mail.So following components I have created.1) Host file deployed on $XXCUST_TOP/bin to send mail.It accepts mail ids,subject and output file.(Code here)2) Concurrent Program to send mail which points to above host file.3) Subscription package to oracle.apps.fnd.concurrent.request.completed.(Code here)Choose a concurrent program which you want to send the out file as attachment.Check Request Completed check box.Submit the program.If it completes normally the business event subscription program will send the out file as attachment to the specified mail id.

    Read the article

  • MYSQL – Identifying Current Version of MySQL Server Installation – Part 2

    - by Pinal Dave
    Earlier I wrote an article about Detecting Current Version of MySQL Server Installation. After the post quite a few emails I received where various users suggested that there are many more ways to figure out the version of MySQL. Here are few of the methods which I received in the email. Method 1: This method retrieves value with the help of Information Functions. SELECT VERSION(); Method 2: This method is very similar to SQL Server. SELECT @@Version Method 3: You can connect to MySQL with command prompt and type following command: STATUS; Method 4: Please refer my earlier blog post. SHOW VARIABLES LIKE "%version%"; Let me know if you know any more method and I will extend this blog post. Reference : Pinal Dave (http://blog.SQLAuthority.com)Filed under: MySQL, PostADay, SQL, SQL Authority, SQL Query, SQL Tips and Tricks, T SQL

    Read the article

  • Could Ajax + Caching be seen as cloaking?

    - by Angel
    I have a website where we use a technique to speed up loading times based in a combination of AJAX + caching. Basically, when we have a section in a page with content which is slow to retrieve, we first look if it's cached. If it is, then we serve the content, if it's not, we serve a placeholder and then make an AJAX call in the client to retrieve the content, wich is now cached for subsequent requests. As a consecuence, sometimes you get the entire page content in the first request, and sometimes you get those placeholders, wich get filled inmediatly with the responses of the AJAX request. You can see an example in the results count by category in the right column of this page: http://www.inzoco.com/crits/2-1-3-28-185-0-28079-0-0/listado-piso-en-alquiler-en-madrid-madrid.aspx I'm worried if it could be seen as cloaking by search engines because if you make a request for a page wich content isn't cached and then ask again for the same page, you would get different responses, the first with the placeholders and AJAX requests and the second one with al the content rendered.

    Read the article

< Previous Page | 183 184 185 186 187 188 189 190 191 192 193 194  | Next Page >