Search Results

Search found 28841 results on 1154 pages for 'simple'.

facebook twitter digg google delicious technorati stumbleupon myspace wordpress linkedin gmail igoogle windows live tumblr viadeo yahoo buzz yahoo mail yahoo bookmarks favorites email print

Page 101/1154 | < Previous Page | 97 98 99 100 101 102 103 104 105 106 107 108  | Next Page >

  • Slow queries in Rails- not sure if my indexes are being used.

    - by Max Williams
    I'm doing a quite complicated find with lots of includes, which rails is splitting into a sequence of discrete queries rather than do a single big join. The queries are really slow - my dataset isn't massive, with none of the tables having more than a few thousand records. I have indexed all of the fields which are examined in the queries but i'm worried that the indexes aren't helping for some reason: i installed a plugin called "query_reviewer" which looks at the queries used to build a page, and lists problems with them. This states that indexes AREN'T being used, and it features the results of calling 'explain' on the query, which lists various problems. Here's an example find call: Question.paginate(:all, {:page=>1, :include=>[:answers, :quizzes, :subject, {:taggings=>:tag}, {:gradings=>[:age_group, :difficulty]}], :conditions=>["((questions.subject_id = ?) or (questions.subject_id = ? and tags.name = ?))", "1", 19, "English"], :order=>"subjects.name, (gradings.difficulty_id is null), gradings.age_group_id, gradings.difficulty_id", :per_page=>30}) And here are the generated sql queries: SELECT DISTINCT `questions`.id FROM `questions` LEFT OUTER JOIN `taggings` ON `taggings`.taggable_id = `questions`.id AND `taggings`.taggable_type = 'Question' LEFT OUTER JOIN `tags` ON `tags`.id = `taggings`.tag_id LEFT OUTER JOIN `subjects` ON `subjects`.id = `questions`.subject_id LEFT OUTER JOIN `gradings` ON gradings.question_id = questions.id WHERE (((questions.subject_id = '1') or (questions.subject_id = 19 and tags.name = 'English'))) ORDER BY subjects.name, (gradings.difficulty_id is null), gradings.age_group_id, gradings.difficulty_id LIMIT 0, 30 SELECT `questions`.`id` AS t0_r0 <..etc...> FROM `questions` LEFT OUTER JOIN `answers` ON answers.question_id = questions.id LEFT OUTER JOIN `quiz_questions` ON (`questions`.`id` = `quiz_questions`.`question_id`) LEFT OUTER JOIN `quizzes` ON (`quizzes`.`id` = `quiz_questions`.`quiz_id`) LEFT OUTER JOIN `subjects` ON `subjects`.id = `questions`.subject_id LEFT OUTER JOIN `taggings` ON `taggings`.taggable_id = `questions`.id AND `taggings`.taggable_type = 'Question' LEFT OUTER JOIN `tags` ON `tags`.id = `taggings`.tag_id LEFT OUTER JOIN `gradings` ON gradings.question_id = questions.id LEFT OUTER JOIN `age_groups` ON `age_groups`.id = `gradings`.age_group_id LEFT OUTER JOIN `difficulties` ON `difficulties`.id = `gradings`.difficulty_id WHERE (((questions.subject_id = '1') or (questions.subject_id = 19 and tags.name = 'English'))) AND `questions`.id IN (602, 634, 666, 698, 730, 762, 613, 645, 677, 709, 741, 592, 624, 656, 688, 720, 752, 603, 635, 667, 699, 731, 763, 614, 646, 678, 710, 742, 593, 625) ORDER BY subjects.name, (gradings.difficulty_id is null), gradings.age_group_id, gradings.difficulty_id SELECT count(DISTINCT `questions`.id) AS count_all FROM `questions` LEFT OUTER JOIN `answers` ON answers.question_id = questions.id LEFT OUTER JOIN `quiz_questions` ON (`questions`.`id` = `quiz_questions`.`question_id`) LEFT OUTER JOIN `quizzes` ON (`quizzes`.`id` = `quiz_questions`.`quiz_id`) LEFT OUTER JOIN `subjects` ON `subjects`.id = `questions`.subject_id LEFT OUTER JOIN `taggings` ON `taggings`.taggable_id = `questions`.id AND `taggings`.taggable_type = 'Question' LEFT OUTER JOIN `tags` ON `tags`.id = `taggings`.tag_id LEFT OUTER JOIN `gradings` ON gradings.question_id = questions.id LEFT OUTER JOIN `age_groups` ON `age_groups`.id = `gradings`.age_group_id LEFT OUTER JOIN `difficulties` ON `difficulties`.id = `gradings`.difficulty_id WHERE (((questions.subject_id = '1') or (questions.subject_id = 19 and tags.name = 'English'))) Actually, looking at these all nicely formatted here, there's a crazy amount of joining going on here. This can't be optimal surely. Anyway, it looks like i have two questions. 1) I have an index on each of the ids and foreign key fields referred to here. The second of the above queries is the slowest, and calling explain on it (doing it directly in mysql) gives me the following: +----+-------------+----------------+--------+---------------------------------------------------------------------------------+-------------------------------------------------+---------+------------------------------------------------+------+----------------------------------------------+ | id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra | +----+-------------+----------------+--------+---------------------------------------------------------------------------------+-------------------------------------------------+---------+------------------------------------------------+------+----------------------------------------------+ | 1 | SIMPLE | questions | range | PRIMARY,index_questions_on_subject_id | PRIMARY | 4 | NULL | 30 | Using where; Using temporary; Using filesort | | 1 | SIMPLE | answers | ref | index_answers_on_question_id | index_answers_on_question_id | 5 | millionaire_development.questions.id | 2 | | | 1 | SIMPLE | quiz_questions | ref | index_quiz_questions_on_question_id | index_quiz_questions_on_question_id | 5 | millionaire_development.questions.id | 1 | | | 1 | SIMPLE | quizzes | eq_ref | PRIMARY | PRIMARY | 4 | millionaire_development.quiz_questions.quiz_id | 1 | | | 1 | SIMPLE | subjects | eq_ref | PRIMARY | PRIMARY | 4 | millionaire_development.questions.subject_id | 1 | | | 1 | SIMPLE | taggings | ref | index_taggings_on_taggable_id_and_taggable_type,index_taggings_on_taggable_type | index_taggings_on_taggable_id_and_taggable_type | 263 | millionaire_development.questions.id,const | 1 | | | 1 | SIMPLE | tags | eq_ref | PRIMARY | PRIMARY | 4 | millionaire_development.taggings.tag_id | 1 | Using where | | 1 | SIMPLE | gradings | ref | index_gradings_on_question_id | index_gradings_on_question_id | 5 | millionaire_development.questions.id | 2 | | | 1 | SIMPLE | age_groups | eq_ref | PRIMARY | PRIMARY | 4 | millionaire_development.gradings.age_group_id | 1 | | | 1 | SIMPLE | difficulties | eq_ref | PRIMARY | PRIMARY | 4 | millionaire_development.gradings.difficulty_id | 1 | | +----+-------------+----------------+--------+---------------------------------------------------------------------------------+-------------------------------------------------+---------+------------------------------------------------+------+----------------------------------------------+ The query_reviewer plugin has this to say about it - it lists several problems: Table questions: Using temporary table, Long key length (263), Using filesort MySQL must do an extra pass to find out how to retrieve the rows in sorted order. To resolve the query, MySQL needs to create a temporary table to hold the result. The key used for the index was rather long, potentially affecting indices in memory 2) It looks like rails isn't splitting this find up in a very optimal way. Is it, do you think? Am i better off doing several find queries manually rather than one big combined one? Grateful for any advice, max

    Read the article

  • On Her Majesty's Secret Source Code: .NET Reflector 7 Early Access Builds Now Available

    - by Bart Read
    Dodgy Bond references aside, I'm extremely happy to be able to tell you that we've just released our first .NET Reflector 7 Early Access build. We're going to make these available over the coming weeks via the main .NET Reflector download page at: http://reflector.red-gate.com/Download.aspx Please have a play and tell us what you think in the forum we've set up. Also, please let us know if you run into any problems in the same place. The new version so far comes with numerous decompilation improvements including (after 5 years!) support for iterator blocks - i.e., the yield statement first seen in .NET 2.0. We've also done a lot of work to solidify the support for .NET 4.0. Clive's written about the work he's done to support iterator blocks in much more detail here, along with the odd problem he's encountered when dealing with compiler generated code: http://www.simple-talk.com/community/blogs/clivet/96199.aspx. On the UI front we've started what will ultimately be a rewrite of the entire front-end, albeit broken into stages over two or three major releases. The most obvious addition at the moment is tabbed browsing, which you can see in Figure 1. Figure 1. .NET Reflector's new tabbed decompilation feature. Use CTRL+Click on any item in the assembly browser tree, or any link in the source code view, to open it in a new tab. This isn't by any means finished. I'll be tying up loose ends for the next few weeks, with a major focus on performance and resource usage. .NET Reflector has historically been a largely single-threaded application which has been fine up until now but, as you might expect, the addition of browser-style tabbing has pushed this approach somewhat beyond its limit. You can see this if you refresh the assemblies list by hitting F5. This shows up another problem: we really need to make Reflector remember everything you had open before you refreshed the list, rather than just the last item you viewed - I discovered that it's always done the latter, but it used to hide all panes apart from the treeview after a Refresh, including the decompiler/disassembler window. Ultimately I've got plans to add the whole VS/Chrome/Firefox style ability to drag a tab into the middle of nowhere to spawn a new window, but I need to be mindful of the add-ins, amongst other things, so it's possible that might slip to a 7.5 or 8.0 release. You'll also notice that .NET Reflector 7 now needs .NET 3.5 or later to run. We made this jump because we wanted to offer ourselves a much better chance of adding some really cool functionality to support newer technologies, such as Silverlight and Windows Phone 7. We've also taken the opportunity to start using WPF for UI development, which has frankly been a godsend. The learning curve is practically vertical but, I kid you not, it's just a far better world. Really. Stop using WinForms. Now. Why are you still using it? I had to go back and work on an old WinForms dialog for an hour or two yesterday and it really made me wince. The point is we'll be able to move the UI in some exciting new directions that will make Reflector easier to use whilst continuing to develop its functionality without (and this is key) cluttering the interface. The 3.5 language enhancements should also enable us to be much more productive over the longer term. I know most of you have .NET Fx 3.5 or 4.0 already but, if you do need to install a new version, I'd recommend you jump straight to 4.0 because, for one thing, it's faster, and if you're starting afresh there's really no reason not to. Despite the Fx version jump the Visual Studio add-in should still work fine in Visual Studio 2005, and obviously will continue to work in Visual Studio 2008 and 2010. If you do run into problems, again, please let us know here. As before, we continue to support every edition of Visual Studio exception the Express Editions. Speaking of Visual Studio, we've also been improving the add-in. You can now open and explore decompiled code for any referenced assembly in any project in your solution. Just right-click on the reference, then click Decompile and Explore on the context menu. Reflector will pop up a progress box whilst it decompiles your assembly (Figure 2) - you can move this out of the way whilst you carry on working. Figure 2. Decompilation progress. This isn't modal so you can just move it out of the way and carry on working. Once it's done you can explore your assembly in the Reflector treeview (Figure 3), also accessible via the .NET Reflector Explore Decompiled Assemblies main menu item. Double-click on any item to open decompiled source in the Visual Studio source code view. Use right-click and Go To Definition on the source view context menu to navigate through the code. Figure 3. Using the .NET Reflector treeview within Visual Studio. Double-click on any item to open decompiled source in the source code view. There are loads of other changes and fixes that have gone in, often under the hood, which I don't have room to talk about here, and plenty more to come over the next few weeks. I'll try to keep you abreast of new functionality and changes as they go in. There are a couple of smaller things worth mentioning now though. Firstly, we've reorganised the menus and toolbar in Reflector itself to more closely mirror what you might be used to in other applications. Secondly, we've tried to make some of the functionality more discoverable. For example, you can now switch decompilation target framework version directly from the toolbar - and the default is now .NET 4.0. I think that about covers it for the moment. As I said, please use the new version, and send us your feedback. Here's that download URL again: http://reflector.red-gate.com/Download.aspx. Until next time! Technorati Tags: .net reflector,7,early access,new version,decompilation,tabbing,visual studio,software development,.net,c#,vb

    Read the article

  • How to get SQL Railroad Diagrams from MSDN BNF syntax notation.

    - by Phil Factor
    pre {margin-bottom:.0001pt; font-size:8.0pt; font-family:"Courier New"; margin-left: 0cm; margin-right: 0cm; margin-top: 0cm; } On SQL Server Books-On-Line, in the Transact-SQL Reference (database Engine), every SQL Statement has its syntax represented in  ‘Backus–Naur Form’ notation (BNF)  syntax. For a programmer in a hurry, this should be ideal because It is the only quick way to understand and appreciate all the permutations of the syntax. It is a great feature once you get your eye in. It isn’t the only way to get the information;  You can, of course, reverse-engineer an understanding of the syntax from the examples, but your understanding won’t be complete, and you’ll have wasted time doing it. BNF is a good start in representing the syntax:  Oracle and SQLite go one step further, and have proper railroad diagrams for their syntax, which is a far more accessible way of doing it. There are three problems with the BNF on MSDN. Firstly, it is isn’t a standard version of  BNF, but an ancient fork from EBNF, inherited from Sybase. Secondly, it is excruciatingly difficult to understand, and thirdly it has a number of syntactic and semantic errors. The page describing DML triggers, for example, currently has the absurd BNF error that makes it state that all statements in the body of the trigger must be separated by commas.  There are a few other detail problems too. Here is the offending syntax for a DML trigger, pasted from MSDN. Trigger on an INSERT, UPDATE, or DELETE statement to a table or view (DML Trigger) CREATE TRIGGER [ schema_name . ]trigger_name ON { table | view } [ WITH <dml_trigger_option> [ ,...n ] ] { FOR | AFTER | INSTEAD OF } { [ INSERT ] [ , ] [ UPDATE ] [ , ] [ DELETE ] } [ NOT FOR REPLICATION ] AS { sql_statement [ ; ] [ ,...n ] | EXTERNAL NAME <method specifier [ ; ] > }   <dml_trigger_option> ::=     [ ENCRYPTION ]     [ EXECUTE AS Clause ]   <method_specifier> ::=  This should, of course, be /* Trigger on an INSERT, UPDATE, or DELETE statement to a table or view (DML Trigger) */ CREATE TRIGGER [ schema_name . ]trigger_name ON { table | view } [ WITH <dml_trigger_option> [ ,...n ] ] { FOR | AFTER | INSTEAD OF } { [ INSERT ] [ , ] [ UPDATE ] [ , ] [ DELETE ] } [ NOT FOR REPLICATION ] AS { {sql_statement [ ; ]} [ ...n ] | EXTERNAL NAME <method_specifier> [ ; ] }   <dml_trigger_option> ::=     [ ENCRYPTION ]     [ EXECUTE AS CLAUSE ]   <method_specifier> ::=     assembly_name.class_name.method_name I’d love to tell Microsoft when I spot errors like this so they can correct them but I can’t. Obviously, there is a mechanism on MSDN to get errors corrected by using comments, but that doesn’t work for me (*Error occurred while saving your data.”), and when I report that the comment system doesn’t work to MSDN, I get no reply. I’ve been trying to create railroad diagrams for all the important SQL Server SQL statements, as good as you’d find for Oracle, and have so far published the CREATE TABLE and ALTER TABLE railroad diagrams based on the BNF. Although I’ve been aware of them, I’ve never realised until recently how many errors there are. Then, Colin Daley created a translator for the SQL Server dialect of  BNF which outputs standard EBNF notation used by the W3C. The example MSDN BNF for the trigger would be rendered as … /* Trigger on an INSERT, UPDATE, or DELETE statement to a table or view (DML Trigger) */ create_trigger ::= 'CREATE TRIGGER' ( schema_name '.' ) ? trigger_name 'ON' ( table | view ) ( 'WITH' dml_trigger_option ( ',' dml_trigger_option ) * ) ? ( 'FOR' | 'AFTER' | 'INSTEAD OF' ) ( ( 'INSERT' ) ? ( ',' ) ? ( 'UPDATE' ) ? ( ',' ) ? ( 'DELETE' ) ? ) ( 'NOT FOR REPLICATION' ) ? 'AS' ( ( sql_statement ( ';' ) ? ) + | 'EXTERNAL NAME' method_specifier ( ';' ) ? )   dml_trigger_option ::= ( 'ENCRYPTION' ) ? ( 'EXECUTE AS CLAUSE' ) ?   method_specifier ::= assembly_name '.' class_name '.' method_name Colin’s intention was to allow anyone to paste SQL Server’s BNF notation into his website-based parser, and from this generate classic railroad diagrams via Gunther Rademacher's Railroad Diagram Generator.  Colin's application does this for you: you're not aware that you are moving to a different site.  Because Colin's 'translator' it is a parser, it will pick up syntax errors. Once you’ve fixed the syntax errors, you will get the syntax in the form of a human-readable railroad diagram and, in this form, the semantic mistakes become flamingly obvious. Gunter’s Railroad Diagram Generator is brilliant. To be able, after correcting the MSDN dialect of BNF, to generate a standard EBNF, and from thence to create railroad diagrams for SQL Server’s syntax that are as good as Oracle’s, is a great boon, and many thanks to Colin for the idea. Here is the result of the W3C EBNF from Colin’s application then being run through the Railroad diagram generator. create_trigger: dml_trigger_option: method_specifier:   Now that’s much better, you’ll agree. This is pretty easy to understand, and at this point any error is immediately obvious. This should be seriously useful, and it is to me. However  there is that snag. The BNF is generally incorrect, and you can’t expect the average visitor to mess about with it. The answer is, of course, to correct the BNF on MSDN and maybe even add railroad diagrams for the syntax. Stop giggling! I agree it won’t happen. In the meantime, we need to collaboratively store and publish these corrected syntaxes ourselves as we do them. How? GitHub?  SQL Server Central?  Simple-Talk? What should those of us who use the system  do with our corrected EBNF so that anyone can use them without hassle?

    Read the article

  • Listing common SQL Code Smells.

    - by Phil Factor
    Once you’ve done a number of SQL Code-reviews, you’ll know those signs in the code that all might not be well. These ’Code Smells’ are coding styles that don’t directly cause a bug, but are indicators that all is not well with the code. . Kent Beck and Massimo Arnoldi seem to have coined the phrase in the "OnceAndOnlyOnce" page of www.C2.com, where Kent also said that code "wants to be simple". Bad Smells in Code was an essay by Kent Beck and Martin Fowler, published as Chapter 3 of the book ‘Refactoring: Improving the Design of Existing Code’ (ISBN 978-0201485677) Although there are generic code-smells, SQL has its own particular coding habits that will alert the programmer to the need to re-factor what has been written. See Exploring Smelly Code   and Code Deodorants for Code Smells by Nick Harrison for a grounding in Code Smells in C# I’ve always been tempted by the idea of automating a preliminary code-review for SQL. It would be so useful to trawl through code and pick up the various problems, much like the classic ‘Lint’ did for C, and how the Code Metrics plug-in for .NET Reflector by Jonathan 'Peli' de Halleux is used for finding Code Smells in .NET code. The problem is that few of the standard procedural code smells are relevant to SQL, and we need an agreed list of code smells. Merrilll Aldrich made a grand start last year in his blog Top 10 T-SQL Code Smells.However, I'd like to make a start by discovering if there is a general opinion amongst Database developers what the most important SQL Smells are. One can be a bit defensive about code smells. I will cheerfully write very long stored procedures, even though they are frowned on. I’ll use dynamic SQL occasionally. You can only use them as an aid for your own judgment and it is fine to ‘sign them off’ as being appropriate in particular circumstances. Also, whole classes of ‘code smells’ may be irrelevant for a particular database. The use of proprietary SQL, for example, is only a ‘code smell’ if there is a chance that the database will have to be ported to another RDBMS. The use of dynamic SQL is a risk only with certain security models. As the saying goes,  a CodeSmell is a hint of possible bad practice to a pragmatist, but a sure sign of bad practice to a purist. Plamen Ratchev’s wonderful article Ten Common SQL Programming Mistakes lists some of these ‘code smells’ along with out-and-out mistakes, but there are more. The use of nested transactions, for example, isn’t entirely incorrect, even though the database engine ignores all but the outermost: but it does flag up the possibility that the programmer thinks that nested transactions are supported. If anything requires some sort of general agreement, the definition of code smells is one. I’m therefore going to make this Blog ‘dynamic, in that, if anyone twitters a suggestion with a #SQLCodeSmells tag (or sends me a twitter) I’ll update the list here. If you add a comment to the blog with a suggestion of what should be added or removed, I’ll do my best to oblige. In other words, I’ll try to keep this blog up to date. The name against each 'smell' is the name of the person who Twittered me, commented about or who has written about the 'smell'. it does not imply that they were the first ever to think of the smell! Use of deprecated syntax such as *= (Dave Howard) Denormalisation that requires the shredding of the contents of columns. (Merrill Aldrich) Contrived interfaces Use of deprecated datatypes such as TEXT/NTEXT (Dave Howard) Datatype mis-matches in predicates that rely on implicit conversion.(Plamen Ratchev) Using Correlated subqueries instead of a join   (Dave_Levy/ Plamen Ratchev) The use of Hints in queries, especially NOLOCK (Dave Howard /Mike Reigler) Few or No comments. Use of functions in a WHERE clause. (Anil Das) Overuse of scalar UDFs (Dave Howard, Plamen Ratchev) Excessive ‘overloading’ of routines. The use of Exec xp_cmdShell (Merrill Aldrich) Excessive use of brackets. (Dave Levy) Lack of the use of a semicolon to terminate statements Use of non-SARGable functions on indexed columns in predicates (Plamen Ratchev) Duplicated code, or strikingly similar code. Misuse of SELECT * (Plamen Ratchev) Overuse of Cursors (Everyone. Special mention to Dave Levy & Adrian Hills) Overuse of CLR routines when not necessary (Sam Stange) Same column name in different tables with different datatypes. (Ian Stirk) Use of ‘broken’ functions such as ‘ISNUMERIC’ without additional checks. Excessive use of the WHILE loop (Merrill Aldrich) INSERT ... EXEC (Merrill Aldrich) The use of stored procedures where a view is sufficient (Merrill Aldrich) Not using two-part object names (Merrill Aldrich) Using INSERT INTO without specifying the columns and their order (Merrill Aldrich) Full outer joins even when they are not needed. (Plamen Ratchev) Huge stored procedures (hundreds/thousands of lines). Stored procedures that can produce different columns, or order of columns in their results, depending on the inputs. Code that is never used. Complex and nested conditionals WHILE (not done) loops without an error exit. Variable name same as the Datatype Vague identifiers. Storing complex data  or list in a character map, bitmap or XML field User procedures with sp_ prefix (Aaron Bertrand)Views that reference views that reference views that reference views (Aaron Bertrand) Inappropriate use of sql_variant (Neil Hambly) Errors with identity scope using SCOPE_IDENTITY @@IDENTITY or IDENT_CURRENT (Neil Hambly, Aaron Bertrand) Schemas that involve multiple dated copies of the same table instead of partitions (Matt Whitfield-Atlantis UK) Scalar UDFs that do data lookups (poor man's join) (Matt Whitfield-Atlantis UK) Code that allows SQL Injection (Mladen Prajdic) Tables without clustered indexes (Matt Whitfield-Atlantis UK) Use of "SELECT DISTINCT" to mask a join problem (Nick Harrison) Multiple stored procedures with nearly identical implementation. (Nick Harrison) Excessive column aliasing may point to a problem or it could be a mapping implementation. (Nick Harrison) Joining "too many" tables in a query. (Nick Harrison) Stored procedure returning more than one record set. (Nick Harrison) A NOT LIKE condition (Nick Harrison) excessive "OR" conditions. (Nick Harrison) User procedures with sp_ prefix (Aaron Bertrand) Views that reference views that reference views that reference views (Aaron Bertrand) sp_OACreate or anything related to it (Bill Fellows) Prefixing names with tbl_, vw_, fn_, and usp_ ('tibbling') (Jeremiah Peschka) Aliases that go a,b,c,d,e... (Dave Levy/Diane McNurlan) Overweight Queries (e.g. 4 inner joins, 8 left joins, 4 derived tables, 10 subqueries, 8 clustered GUIDs, 2 UDFs, 6 case statements = 1 query) (Robert L Davis) Order by 3,2 (Dave Levy) MultiStatement Table functions which are then filtered 'Sel * from Udf() where Udf.Col = Something' (Dave Ballantyne) running a SQL 2008 system in SQL 2000 compatibility mode(John Stafford)

    Read the article

  • SQL Monitor’s data repository

    - by Chris Lambrou
    As one of the developers of SQL Monitor, I often get requests passed on by our support people from customers who are looking to dip into SQL Monitor’s own data repository, in order to pull out bits of information that they’re interested in. Since there’s clearly interest out there in playing around directly with the data repository, I thought I’d write some blog posts to start to describe how it all works. The hardest part for me is knowing where to begin, since the schema of the data repository is pretty big. Hmmm… I guess it’s tricky for anyone to write anything but the most trivial of queries against the data repository without understanding the hierarchy of monitored objects, so perhaps my first post should start there. I always imagine that whenever a customer fires up SSMS and starts to explore their SQL Monitor data repository database, they become immediately bewildered by the schema – that was certainly my experience when I did so for the first time. The following query shows the number of different object types in the data repository schema: SELECT type_desc, COUNT(*) AS [count] FROM sys.objects GROUP BY type_desc ORDER BY type_desc;  type_desccount 1DEFAULT_CONSTRAINT63 2FOREIGN_KEY_CONSTRAINT181 3INTERNAL_TABLE3 4PRIMARY_KEY_CONSTRAINT190 5SERVICE_QUEUE3 6SQL_INLINE_TABLE_VALUED_FUNCTION381 7SQL_SCALAR_FUNCTION2 8SQL_STORED_PROCEDURE100 9SYSTEM_TABLE41 10UNIQUE_CONSTRAINT54 11USER_TABLE193 12VIEW124 With 193 tables, 124 views, 100 stored procedures and 381 table valued functions, that’s quite a hefty schema, and when you browse through it using SSMS, it can be a bit daunting at first. So, where to begin? Well, let’s narrow things down a bit and only look at the tables belonging to the data schema. That’s where all of the collected monitoring data is stored by SQL Monitor. The following query gives us the names of those tables: SELECT sch.name + '.' + obj.name AS [name] FROM sys.objects obj JOIN sys.schemas sch ON sch.schema_id = obj.schema_id WHERE obj.type_desc = 'USER_TABLE' AND sch.name = 'data' ORDER BY sch.name, obj.name; This query still returns 110 tables. I won’t show them all here, but let’s have a look at the first few of them:  name 1data.Cluster_Keys 2data.Cluster_Machine_ClockSkew_UnstableSamples 3data.Cluster_Machine_Cluster_StableSamples 4data.Cluster_Machine_Keys 5data.Cluster_Machine_LogicalDisk_Capacity_StableSamples 6data.Cluster_Machine_LogicalDisk_Keys 7data.Cluster_Machine_LogicalDisk_Sightings 8data.Cluster_Machine_LogicalDisk_UnstableSamples 9data.Cluster_Machine_LogicalDisk_Volume_StableSamples 10data.Cluster_Machine_Memory_Capacity_StableSamples 11data.Cluster_Machine_Memory_UnstableSamples 12data.Cluster_Machine_Network_Capacity_StableSamples 13data.Cluster_Machine_Network_Keys 14data.Cluster_Machine_Network_Sightings 15data.Cluster_Machine_Network_UnstableSamples 16data.Cluster_Machine_OperatingSystem_StableSamples 17data.Cluster_Machine_Ping_UnstableSamples 18data.Cluster_Machine_Process_Instances 19data.Cluster_Machine_Process_Keys 20data.Cluster_Machine_Process_Owner_Instances 21data.Cluster_Machine_Process_Sightings 22data.Cluster_Machine_Process_UnstableSamples 23… There are two things I want to draw your attention to: The table names describe a hierarchy of the different types of object that are monitored by SQL Monitor (e.g. clusters, machines and disks). For each object type in the hierarchy, there are multiple tables, ending in the suffixes _Keys, _Sightings, _StableSamples and _UnstableSamples. Not every object type has a table for every suffix, but the _Keys suffix is especially important and a _Keys table does indeed exist for every object type. In fact, if we limit the query to return only those tables ending in _Keys, we reveal the full object hierarchy: SELECT sch.name + '.' + obj.name AS [name] FROM sys.objects obj JOIN sys.schemas sch ON sch.schema_id = obj.schema_id WHERE obj.type_desc = 'USER_TABLE' AND sch.name = 'data' AND obj.name LIKE '%_Keys' ORDER BY sch.name, obj.name;  name 1data.Cluster_Keys 2data.Cluster_Machine_Keys 3data.Cluster_Machine_LogicalDisk_Keys 4data.Cluster_Machine_Network_Keys 5data.Cluster_Machine_Process_Keys 6data.Cluster_Machine_Services_Keys 7data.Cluster_ResourceGroup_Keys 8data.Cluster_ResourceGroup_Resource_Keys 9data.Cluster_SqlServer_Agent_Job_History_Keys 10data.Cluster_SqlServer_Agent_Job_Keys 11data.Cluster_SqlServer_Database_BackupType_Backup_Keys 12data.Cluster_SqlServer_Database_BackupType_Keys 13data.Cluster_SqlServer_Database_CustomMetric_Keys 14data.Cluster_SqlServer_Database_File_Keys 15data.Cluster_SqlServer_Database_Keys 16data.Cluster_SqlServer_Database_Table_Index_Keys 17data.Cluster_SqlServer_Database_Table_Keys 18data.Cluster_SqlServer_Error_Keys 19data.Cluster_SqlServer_Keys 20data.Cluster_SqlServer_Services_Keys 21data.Cluster_SqlServer_SqlProcess_Keys 22data.Cluster_SqlServer_TopQueries_Keys 23data.Cluster_SqlServer_Trace_Keys 24data.Group_Keys The full object type hierarchy looks like this: Cluster Machine LogicalDisk Network Process Services ResourceGroup Resource SqlServer Agent Job History Database BackupType Backup CustomMetric File Table Index Error Services SqlProcess TopQueries Trace Group Okay, but what about the individual objects themselves represented at each level in this hierarchy? Well that’s what the _Keys tables are for. This is probably best illustrated by way of a simple example – how can I query my own data repository to find the databases on my own PC for which monitoring data has been collected? Like this: SELECT clstr._Name AS cluster_name, srvr._Name AS instance_name, db._Name AS database_name FROM data.Cluster_SqlServer_Database_Keys db JOIN data.Cluster_SqlServer_Keys srvr ON db.ParentId = srvr.Id -- Note here how the parent of a Database is a Server JOIN data.Cluster_Keys clstr ON srvr.ParentId = clstr.Id -- Note here how the parent of a Server is a Cluster WHERE clstr._Name = 'dev-chrisl2' -- This is the hostname of my own PC ORDER BY clstr._Name, srvr._Name, db._Name;  cluster_nameinstance_namedatabase_name 1dev-chrisl2SqlMonitorData 2dev-chrisl2master 3dev-chrisl2model 4dev-chrisl2msdb 5dev-chrisl2mssqlsystemresource 6dev-chrisl2tempdb 7dev-chrisl2sql2005SqlMonitorData 8dev-chrisl2sql2005TestDatabase 9dev-chrisl2sql2005master 10dev-chrisl2sql2005model 11dev-chrisl2sql2005msdb 12dev-chrisl2sql2005mssqlsystemresource 13dev-chrisl2sql2005tempdb 14dev-chrisl2sql2008SqlMonitorData 15dev-chrisl2sql2008master 16dev-chrisl2sql2008model 17dev-chrisl2sql2008msdb 18dev-chrisl2sql2008mssqlsystemresource 19dev-chrisl2sql2008tempdb These results show that I have three SQL Server instances on my machine (a default instance, one named sql2005 and one named sql2008), and each instance has the usual set of system databases, along with a database named SqlMonitorData. Basically, this is where I test SQL Monitor on different versions of SQL Server, when I’m developing. There are a few important things we can learn from this query: Each _Keys table has a column named Id. This is the primary key. Each _Keys table has a column named ParentId. A foreign key relationship is defined between each _Keys table and its parent _Keys table in the hierarchy. There are two exceptions to this, Cluster_Keys and Group_Keys, because clusters and groups live at the root level of the object hierarchy. Each _Keys table has a column named _Name. This is used to uniquely identify objects in the table within the scope of the same shared parent object. Actually, that last item isn’t always true. In some cases, the _Name column is actually called something else. For example, the data.Cluster_Machine_Services_Keys table has a column named _ServiceName instead of _Name (sorry for the inconsistency). In other cases, a name isn’t sufficient to uniquely identify an object. For example, right now my PC has multiple processes running, all sharing the same name, Chrome (one for each tab open in my web-browser). In such cases, multiple columns are used to uniquely identify an object within the scope of the same shared parent object. Well, that’s it for now. I’ve given you enough information for you to explore the _Keys tables to see how objects are stored in your own data repositories. In a future post, I’ll try to explain how monitoring data is stored for each object, using the _StableSamples and _UnstableSamples tables. If you have any questions about this post, or suggestions for future posts, just submit them in the comments section below.

    Read the article

  • Database continuous integration step by step

    - by David Atkinson
    This post will describe how to set up basic database continuous integration using TeamCity to initiate the build process, SQL Source Control to put your database under source control, and the SQL Compare command line to keep a test database up to date. In my example I will be using Subversion as my source control repository. If you wish to follow my steps verbatim, please make sure you have TortoiseSVN, SQL Compare and SQL Source Control installed. Downloading and Installing TeamCity TeamCity (http://www.jetbrains.com/teamcity/index.html) is free for up to three agents, so it a great no-risk tool you can use to experiment with. 1. Download the latest version from the JetBrains website. For some reason the TeamCity executable didn't download properly for me, stalling frustratingly at 99%, so I tried again with the zip file download option (see screenshot below), which worked flawlessly. 2. Run the installer using the defaults. This results in a set-up with the server component and agent installed on the same machine, which is ideal for getting started with ease. 3. Check that the build agent is pointing to the server correctly. This has caught me out a few times before. This setting is in C:\TeamCity\buildAgent\conf\buildAgent.properties and for my installation is serverUrl=http\://localhost\:80 . If you need to change this value, if for example you've had to install the Server console to a different port number, the TeamCity Build Agent Service will need to be restarted for the change to take effect. 4. Open the TeamCity admin console on http://localhost , and specify your own designated username and password at first startup. Putting your database in source control using SQL Source Control 5. Assuming you've got SQL Source Control installed, select a development database in the SQL Server Management Studio Object Explorer and select Link Database to Source Control. 6. For the Link step you can either create your own empty folder in source control, or you can select Just Evaluating, which just creates a local subversion repository for you behind the scenes. 7. Once linked, note that your database turns green in the Object Explorer. Visit the Commit tab to do an initial commit of your database objects by typing in an appropriate comment and clicking Commit. 8. There is a hidden feature in SQL Source Control that opens up TortoiseSVN (provided it is installed) pointing to the linked repository. Keep Shift depressed and right click on the text to the right of 'Linked to', in the example below, it's the red Evaluation Repository text. Select Open TortoiseSVN Repo Browser. This screen should give you an idea of how SQL Source Control manages the object files behind the scenes. Back in the TeamCity admin console, we'll now create a new project to monitor the above repository location and to trigger a 'build' each time the repository changes. 9. In TeamCity Adminstration, select Create Project and give it a name, such as "My first database CI", and click Create. 10. Click on Create Build Configuration, and name it something like "Integration build". 11. Click VCS settings and then Create And Attach new VCS root. This is where you will tell TeamCity about the repository it should monitor. 12. In my case since I'm using the Just Evaluating option in SQL Source Control, I should select Subversion. 13. In the URL field paste your repository location. In my case this is file:///C:/Users/David.Atkinson/AppData/Local/Red Gate/SQL Source Control 3/EvaluationRepositories/WidgetDevelopment/WidgetDevelopment 14. Click on Test Connection to ensure that you can communicate with your source control system. Click Save. 15. Click Add Build Step, and Runner Type: Command Line. Should you be familiar with the other runner types, such as NAnt, MSBuild or Powershell, you can opt for these, but for the same of keeping it simple I will pick the simplest option. 16. If you have installed SQL Compare in the default location, set the Command Executable field to: C:\Program Files (x86)\Red Gate\SQL Compare 10\sqlcompare.exe 17. Flip back to SSMS briefly and add a new database to your server. This will be the database used for continuous integration testing. 18. Set the command parameters according to your server and the name of the database you have created. In my case I created database RedGateCI on server .\sql2008r2 /scripts1:. /server2:.\sql2008r2 /db2:RedGateCI /sync /verbose Note that if you pick a server instance that isn't on your local machine, you'll need the TCP/IP protocol enabled in SQL Server Configuration Manager otherwise the SQL Compare command line will not be able to connect. 19. Save and select Build Triggering / Add New Trigger / VCS Trigger. This is where you tell TeamCity when it should initiate a build. Click Save. 20. Now return to SQL Server Management Studio and make a schema change (eg add a new object) to your linked development database. A blue indicator will appear in the Object Explorer. Commit this change, typing in an appropriate check-in comment. All being good, within 60 seconds (a TeamCity default that can be changed) a build will be triggered. 21. Click on Projects in TeamCity to get back to the overview screen: The build log will show you the console output, which is useful for troubleshooting any issues: That's it! You now have continuous integration on your database. In future posts I'll cover how you can generate and test the database creation script, the database upgrade script, and run database unit tests as part of your continuous integration script. If you have any trouble getting this up and running please let me know, either by commenting on this post, or email me directly using the email address below. Technorati Tags: SQL Server

    Read the article

  • Rockmelt, the technology adoption model, and Facebook's spare internet

    - by Roger Hart
    Regardless of how good it is, you'd have to have a heart of stone not to make snide remarks about Rockmelt. After all, on the surface it looks a lot like some people spent two years building a browser instead of just bashing out a Chrome extension over a wet weekend. It probably does some more stuff. I don't know for sure because artificial scarcity is cool, apparently, so the "invitation" is still in the post*. I may in fact never know for sure, because I'm not wild about Facebook sign-in as a prerequisite for anything. From the video, and some initial reviews, my early reaction was: I have a browser, I have a Twitter client; what on earth is this for? The answer, of course, is "not me". Rockmelt is, in a way, quite audacious. Oh, sure, on launch day it's Bay Area bar-chat for the kids with no lenses in their retro specs and trousers that give you deep-vein thrombosis, but it's not really about them. Likewise,  Facebook just launched Google Wave, or something. And all the tech snobbery and scorn packed into describing it that way is irrelevant next to what they're doing with their platform. Here's something I drew in MS Paint** because I don't want to get sued: (see: The technology adoption lifecycle) A while ago in the Guardian, John Lanchester dusted off the idiom that "technology is stuff that doesn't work yet". The rest of the article would be quite interesting if it wasn't largely about MySpace, and he's sort of got a point. If you bolt on the sentiment that risk-averse businessmen like things that work, you've got the essence of Crossing the Chasm. Products for the mainstream market don't look much like technology. Think for  a second about early (1980s ish) hi-fi systems, with all the knobs and fiddly bits, their ostentatious technophile aesthetic. Then consider their sleeker and less (or at least less conspicuously) functional successors in the 1990s. The theory goes that innovators and early adopters like technology, it's a hobby in itself. The rest of the humans seem to like magic boxes with very few buttons that make stuff happen and never trouble them about why. Personally, I consider Apple's maddening insistence that iTunes is an acceptable way to move files around to be more or less morally unacceptable. Most people couldn't care less. Hence Rockmelt, and hence Facebook's continued growth. Rockmelt looks pointless to me, because I aggregate my social gubbins with Digsby, or use TweetDeck. But my use case is different and so are my enthusiasms. If I want to share photos, I'll use Flickr - but Facebook has photo sharing. If I want a short broadcast message, I'll use Twitter - Facebook has status updates. If I want to sell something with relatively little hassle, there's eBay - or Facebook marketplace. YouTube - check, FB Video. Email - messaging. Calendaring apps, yeah there are loads, or FB Events. What if I want to host a simple web page? Sure, they've got pages. Also Notes for blogging, and more games than I can count. This stuff is right there, where millions and millions of users are already, and for what they need it just works. It's not about me, because I'm not in the big juicy area under the curve. It's what 1990s portal sites could never have dreamed of achieving. Facebook is AOL on speed, crack, and some designer drugs it had specially imported from the future. It's a n00b-friendly gateway to the internet that just happens to serve up all the things you want to do online, right where you are. Oh, and everybody else is there too. The price of having all this and the social graph too is that you have all of this, and the social graph too. But plenty of folks have more incisive things to say than me about the whole privacy shebang, and it's not really what I'm talking about. Facebook is maintaining a vast, and fairly fully-featured training-wheels internet. And it makes up a large proportion of the online experience for a lot of people***. It's the entire web (2.0?) experience for the early and late majority. And sure, no individual bit of it is quite as slick or as fully-realised as something like Flickr (which wows me a bit every time I use it. Those guys are good at the web), but it doesn't have to be. It has to be unobtrusively good enough for the regular humans. It has to not feel like technology. This is what Rockmelt sort of is. You're online, you want something nebulously social, and you don't want to faff about with, say, Twitter clients. Wow! There it is on a really distracting sidebar, right in your browser. No effort! Yeah - fish nor fowl, much? It might work, I guess. There may be a demographic who want their social web experience more simply than tech tinkering, and who aren't just getting it from Facebook (or, for that matter, mobile devices). But I'd be surprised. Rockmelt feels like an attempt to grab a slice of Facebook-style "Look! It's right here, where you already are!", but it's still asking the mature market to install a new browser. Presumably this is where that Facebook sign-in predicate comes in handy, though it'll take some potent awareness marketing to make it fly. Meanwhile, Facebook quietly has the entire rest of the internet as a product management resource, and can continue to give most of the people most of what they want. Something that has not gone un-noticed in its potential to look a little sinister. But heck, they might even make Google Wave popular.     *This was true last week when I drafted this post. I got an invite subsequently, hence the screenshot. **MS Paint is no fun any more. It's actually good in Windows 7. Farewell ironically-shonky diagrams. *** It's also behind a single sign-in, lending a veneer of confidence, and partially solving the problem of usernames being crummy unique identifiers. I'll be blogging about that at some point.

    Read the article

  • .NET to iOS: From WinForms to the iPad

    - by RobertChipperfield
    One of the great things about working at Red Gate is getting to play with new technology - and right now, that means mobile. A few weeks ago, we decided that a little research into the tablet computing arena was due, and purely from a numbers point of view, that suggested the iPad as a good target device. A quick trip to iPhoneDevCon in San Diego later, and Marine and I came back full of ideas, and with some concept of how iOS development was meant to work. Here's how we went from there to the release of Stacks & Heaps, our geeky take on the classic "Snakes & Ladders" game. Step 1: Buy a Mac I've played with many operating systems in my time: from the original BBC Model B, through DOS, Windows, Linux, and others, but I'd so far managed to avoid buying fruit-flavoured computer hardware! If you want to develop for the iPhone, iPad or iPod Touch, that's the first thing that needs to change. If you've not used OS X before, the first thing you'll realise is that everything is different! In the interests of avoiding a flame war in the comments section, I'll only go so far as to say that a lot of my Windows-flavoured muscle memory no longer worked. If you're in the UK, you'll also realise your keyboard is lacking a # key, and that " and @ are the other way around from normal. The wonderful Ukelele keyboard layout editor restores some sanity here, as long as you don't look at the keyboard when you're typing. I couldn't give up the PC entirely, but a handy application called Synergy comes to the rescue - it lets you share a single keyboard and mouse between multiple machines. There's a few limitations: Alt-Tab always seems to go to the Mac, and Windows 7's UAC dialogs require the local mouse for security reasons, but it gets you a long way at least. Step 2: Register as an Apple Developer You can register as an Apple Developer free of charge, and that lets you download XCode and the iOS SDK. You also get the iPhone / iPad emulator, which is handy, since you'll need to be a paid member before you can deploy your apps to a real device. You can either enroll as an individual, or as a company. They both cost the same ($99/year), but there's a few differences between them. If you register as a company, you can add multiple developers to your team (all for the same $99 - not $99 per developer), and you get to use your company name in the App Store. However, you'll need to send off significantly more documentation to Apple, and I suspect the process takes rather longer than for an individual, where they just need to verify some credit card details. Here's a tip: if you're registering as a company, do so as early as possible. The approval process can take a while to complete, so get the application in in plenty of time. Step 3: Learn to love the square brackets! Objective-C is the language of the iPad. C and C++ are also supported, and if you're doing some serious game development, you'll probably spend most of your time in C++ talking OpenGL, but for forms-based apps, you'll be interacting with a lot of the Objective-C SDK. Like shifting from Ctrl-C to Cmd-C, it feels a little odd at first, with the familiar string.format(.) turning into: NSString *myString = [NSString stringWithFormat:@"Hello world, it's %@", [NSDate date]]; Thankfully XCode's auto-complete is normally passable, if not up to Visual Studio's standards, which coupled with a huge amount of content on Stack Overflow means you'll soon get to grips with the API. You'll need to get used to some terminology changes, though; here's an incomplete approximation: Coming from a .NET background, there's some luxuries you no longer have developing Objective C in XCode: Generics! Remember back in .NET 1.1, when all collections were just objects? Yup, we're back there now. ReSharper. Or, more generally, very much refactoring support. The not-many-keystrokes to rename a class, its file, and al references to it in Visual Studio turns into a much more painful experience in XCode. Garbage collection. This is actually rather less of an issue than you might expect: if you follow the rules, the reference counting provided by Objective C gets you a long way without too much pain. Circular references are their usual problematic self, though. Decent exception handling. You do have exceptions, but they're nowhere near as widely used. Generally, if something goes wrong, you get nil (see translation table above) back. Which brings me on to. Calling a method on a nil object isn't a failure - it just returns nil itself! There's many arguments for and against this, but personally I fall into the "stuff should fail as quickly and explicitly as possible" camp. Less specifically, I found that there's more chance of code failing at runtime rather than getting caught at compile-time: using the @selector(.) syntax to pass a method signature isn't (can't be) checked at compile-time, so the first you know about a typo is a crash when you try and call it. The solution to this is of course lots of great testing, both automated and manual, but I still find comfort in provably correct type safety being enforced in addition to testing. Step 4: Submit to the App Store Assuming you want to distribute to more than a handful of devices, you're going to need to submit your app to the Apple App Store. There's a few gotchas in terms of getting builds signed with the right certificates, and you'll be bouncing around between XCode and iTunes Connect a fair bit, but eventually you get everything checked off the to-do list, and are ready to upload your first binary! With some amount of anticipation, I pressed the Upload button in XCode, ready to release our creation into the world, but was instead greeted by an error informing me my XML file was malformed. Uh. A little Googling later, and it turned out that a simple rename from "Stacks&Heaps.app" to "StacksAndHeaps.app" worked around an XML escaping bug, and we were good to go. The next step is to wait for approval (or otherwise). After a couple of weeks of intensive development, this part is agonising. Did we make it? The Apple jury is still out at the moment, but our fingers are firmly crossed! In the meantime, you can see some screenshots and leave us your email address if you'd like us to get in touch when it does go live at the MobileFoo website. Step 5: Profit! Actually, that wasn't the idea here: Stacks & Heaps is free; there's no adverts, and we're not going to sell all your data either. So why did we do it? We wanted to get an idea of what it's like to move from coding for a desktop environment, to something completely different. We don't know whether in a year's time, the iPad will still be the dominant force, or whether Android will have smoothed out some bugs, tweaked the performance, and polished the UI, but I think it's a fairly sure bet that the tablet form factor is here to stay. We want to meet people who are using it, start chatting to them, and find out about some of the pain they're feeling. What better way to do that than do it ourselves, and get to write a cool game in the process?

    Read the article

  • Working With Extended Events

    - by Fatherjack
    SQL Server 2012 has made working with Extended Events (XE) pretty simple when it comes to what sessions you have on your servers and what options you have selected and so forth but if you are like me then you still have some SQL Server instances that are 2008 or 2008 R2. For those servers there is no built-in way to view the Extended Event sessions in SSMS. I keep coming up against the same situations – Where are the xel log files? What events, actions or predicates are set for the events on the server? What sessions are there on the server already? I got tired of this being a perpetual question and wrote some TSQL to save as a snippet in SQL Prompt so that these details are permanently only a couple of clicks away. First, some history. If you just came here for the code skip down a few paragraphs and it’s all there. If you want a little time to reminisce about SQL Server then stick with me through the next paragraph or two. We are in a bit of a cross-over period currently, there are many versions of SQL Server but I would guess that SQL Server 2008, 2008 R2 and 2012 comprise the majority of installations. With each of these comes a set of management tools, of which SQL Server Management Studio (SSMS) is one. In 2008 and 2008 R2 Extended Events made their first appearance and there was no way to work with them in the SSMS interface. At some point the Extended Events guru Jonathan Kehayias (http://www.sqlskills.com/blogs/jonathan/) created the SQL Server 2008 Extended Events SSMS Addin which is really an excellent tool to ease XE session administration. This addin will install in SSMS 2008 or 2008R2 but not SSMS 2012. If you use a compatible version of SSMS then I wholly recommend downloading and using it to make your work with XE much easier. If you have SSMS 2012 installed, and there is no reason not to as it will let you work with all versions of SQL Server, then you cannot install this addin. If you are working with SQL Server 2012 then SSMS 2012 has built in functionality to manage XE sessions – this functionality does not apply for 2008 or 2008 R2 instances though. This means you are somewhat restricted and have to use TSQL to manage XE sessions on older versions of SQL Server. OK, those of you that skipped ahead for the code, you need to start from here: So, you are working with SSMS 2012 but have a SQL Server that is an earlier version that needs an XE session created or you think there is a session created but you aren’t sure, or you know it’s there but can’t remember if it is running and where the output is going. How do you find out? Well, none of the information is hidden as such but it is a bit of a wrangle to locate it and it isn’t a lot of code that is unlikely to remain in your memory. I have created two pieces of code. The first examines the SYS.Server_Event_… management views in combination with the SYS.DM_XE_… management views to give the name of all sessions that exist on the server, regardless of whether they are running or not and two pieces of TSQL code. One piece will alter the state of the session: if the session is running then the code will stop the session if executed and vice versa. The other piece of code will drop the selected session. If the session is running then the code will stop it first. Do not execute the DROP code unless you are sure you have the Create code to hand. It will be dropped from the server without a second chance to change your mind. /**************************************************************/ /***   To locate and describe event sessions on a server    ***/ /***                                                        ***/ /***   Generates TSQL to start/stop/drop sessions           ***/ /***                                                        ***/ /***        Jonathan Allen - @fatherjack                    ***/ /***                 June 2013                                ***/ /***                                                        ***/ /**************************************************************/ SELECT  [EES].[name] AS [Session Name - all sessions] ,         CASE WHEN [MXS].[name] IS NULL THEN ISNULL([MXS].[name], 'Stopped')              ELSE 'Running'         END AS SessionState ,         CASE WHEN [MXS].[name] IS NULL              THEN ISNULL([MXS].[name],                          'ALTER EVENT SESSION [' + [EES].[name]                          + '] ON SERVER STATE = START;')              ELSE 'ALTER EVENT SESSION [' + [EES].[name]                   + '] ON SERVER STATE = STOP;'         END AS ALTER_SessionState ,         CASE WHEN [MXS].[name] IS NULL              THEN ISNULL([MXS].[name],                          'DROP EVENT SESSION [' + [EES].[name]                          + '] ON SERVER; -- This WILL drop the session. It will no longer exist. Don't do it unless you are certain you can recreate it if you need it.')              ELSE 'ALTER EVENT SESSION [' + [EES].[name]                   + '] ON SERVER STATE = STOP; ' + CHAR(10)                   + '-- DROP EVENT SESSION [' + [EES].[name]                   + '] ON SERVER; -- This WILL stop and drop the session. It will no longer exist. Don't do it unless you are certain you can recreate it if you need it.'         END AS DROP_Session FROM    [sys].[server_event_sessions] AS EES         LEFT JOIN [sys].[dm_xe_sessions] AS MXS ON [EES].[name] = [MXS].[name] WHERE   [EES].[name] NOT IN ( 'system_health', 'AlwaysOn_health' ) ORDER BY SessionState GO I have excluded the system_health and AlwaysOn sessions as I don’t want to accidentally execute the drop script for these sessions that are created as part of the SQL Server installation. It is possible to recreate the sessions but that is a whole lot of aggravation I’d rather avoid. The second piece of code gathers details of running XE sessions only and provides information on the Events being collected, any predicates that are set on those events, the actions that are set to be collected, where the collected information is being logged and if that logging is to a file target, where that file is located. /**********************************************/ /***    Running Session summary                ***/ /***                                        ***/ /***    Details key values of XE sessions     ***/ /***    that are in a running state            ***/ /***                                        ***/ /***        Jonathan Allen - @fatherjack    ***/ /***        June 2013                        ***/ /***                                        ***/ /**********************************************/ SELECT  [EES].[name] AS [Session Name - running sessions] ,         [EESE].[name] AS [Event Name] ,         COALESCE([EESE].[predicate], 'unfiltered') AS [Event Predicate Filter(s)] ,         [EESA].[Action] AS [Event Action(s)] ,         [EEST].[Target] AS [Session Target(s)] ,         ISNULL([EESF].[value], 'No file target in use') AS [File_Target_UNC] -- select * FROM    [sys].[server_event_sessions] AS EES         INNER JOIN [sys].[dm_xe_sessions] AS MXS ON [EES].[name] = [MXS].[name]         INNER JOIN [sys].[server_event_session_events] AS [EESE] ON [EES].[event_session_id] = [EESE].[event_session_id]         LEFT JOIN [sys].[server_event_session_fields] AS EESF ON ( [EES].[event_session_id] = [EESF].[event_session_id]                                                               AND [EESF].[name] = 'filename'                                                               )         CROSS APPLY ( SELECT    STUFF(( SELECT  ', ' + sest.name                                         FROM    [sys].[server_event_session_targets]                                                 AS SEST                                         WHERE   [EES].[event_session_id] = [SEST].[event_session_id]                                       FOR                                         XML PATH('')                                       ), 1, 2, '') AS [Target]                     ) AS EEST         CROSS APPLY ( SELECT    STUFF(( SELECT  ', ' + [sesa].NAME                                         FROM    [sys].[server_event_session_actions]                                                 AS sesa                                         WHERE   [sesa].[event_session_id] = [EES].[event_session_id]                                       FOR                                         XML PATH('')                                       ), 1, 2, '') AS [Action]                     ) AS EESA WHERE   [EES].[name] NOT IN ( 'system_health', 'AlwaysOn_health' ) /*Optional to exclude 'out-of-the-box' traces*/ I hope that these scripts are useful to you and I would be obliged if you would keep my name in the script comments. I have no problem with you using it in production or personal circumstances, however it has no warranty or guarantee. Don’t use it unless you understand it and are happy with what it is going to do. I am not ever responsible for the consequences of executing this script on your servers.

    Read the article

  • Anatomy of a .NET Assembly - Custom attribute encoding

    - by Simon Cooper
    In my previous post, I covered how field, method, and other types of signatures are encoded in a .NET assembly. Custom attribute signatures differ quite a bit from these, which consequently affects attribute specifications in C#. Custom attribute specifications In C#, you can apply a custom attribute to a type or type member, specifying a constructor as well as the values of fields or properties on the attribute type: public class ExampleAttribute : Attribute { public ExampleAttribute(int ctorArg1, string ctorArg2) { ... } public Type ExampleType { get; set; } } [Example(5, "6", ExampleType = typeof(string))] public class C { ... } How does this specification actually get encoded and stored in an assembly? Specification blob values Custom attribute specification signatures use the same building blocks as other types of signatures; the ELEMENT_TYPE structure. However, they significantly differ from other types of signatures, in that the actual parameter values need to be stored along with type information. There are two types of specification arguments in a signature blob; fixed args and named args. Fixed args are the arguments to the attribute type constructor, named arguments are specified after the constructor arguments to provide a value to a field or property on the constructed attribute type (PropertyName = propValue) Values in an attribute blob are limited to one of the basic types (one of the number types, character, or boolean), a reference to a type, an enum (which, in .NET, has to use one of the integer types as a base representation), or arrays of any of those. Enums and the basic types are easy to store in a blob - you simply store the binary representation. Strings are stored starting with a compressed integer indicating the length of the string, followed by the UTF8 characters. Array values start with an integer indicating the number of elements in the array, then the item values concatentated together. Rather than using a coded token, Type values are stored using a string representing the type name and fully qualified assembly name (for example, MyNs.MyType, MyAssembly, Version=1.0.0.0, Culture=neutral, PublicKeyToken=0123456789abcdef). If the type is in the current assembly or mscorlib then just the type name can be used. This is probably done to prevent direct references between assemblies solely because of attribute specification arguments; assemblies can be loaded in the reflection-only context and attribute arguments still processed, without loading the entire assembly. Fixed and named arguments Each entry in the CustomAttribute metadata table contains a reference to the object the attribute is applied to, the attribute constructor, and the specification blob. The number and type of arguments to the constructor (the fixed args) can be worked out by the method signature referenced by the attribute constructor, and so the fixed args can simply be concatenated together in the blob without any extra type information. Named args are different. These specify the value to assign to a field or property once the attribute type has been constructed. In the CLR, fields and properties can be overloaded just on their type; different fields and properties can have the same name. Therefore, to uniquely identify a field or property you need: Whether it's a field or property (indicated using byte values 0x53 and 0x54, respectively) The field or property type The field or property name After the fixed arg values is a 2-byte number specifying the number of named args in the blob. Each named argument has the above information concatenated together, mostly using the basic ELEMENT_TYPE values, in the same way as a method or field signature. A Type argument is represented using the byte 0x50, and an enum argument is represented using the byte 0x55 followed by a string specifying the name and assembly of the enum type. The named argument property information is followed by the argument value, using the same encoding as fixed args. Boxed objects This would be all very well, were it not for object and object[]. Arguments and properties of type object allow a value of any allowed argument type to be specified. As a result, more information needs to be specified in the blob to interpret the argument bytes as the correct type. So, the argument value is simple prepended with the type of the value by specifying the ELEMENT_TYPE or name of the enum the value represents. For named arguments, a field or property of type object is represented using the byte 0x51, with the actual type specified in the argument value. Some examples... All property signatures start with the 2-byte value 0x0001. Similar to my previous post in the series, names in capitals correspond to a particular byte value in the ELEMENT_TYPE structure. For strings, I'll simply give the string value, rather than the length and UTF8 encoding in the actual blob. I'll be using the following enum and attribute types to demonstrate specification encodings: class AttrAttribute : Attribute { public AttrAttribute() {} public AttrAttribute(Type[] tArray) {} public AttrAttribute(object o) {} public AttrAttribute(MyEnum e) {} public AttrAttribute(ushort x, int y) {} public AttrAttribute(string str, Type type1, Type type2) {} public int Prop1 { get; set; } public object Prop2 { get; set; } public object[] ObjectArray; } enum MyEnum : int { Val1 = 1, Val2 = 2 } Now, some examples: Here, the the specification binds to the (ushort, int) attribute constructor, with fixed args only. The specification blob starts off with a prolog, followed by the two constructor arguments, then the number of named arguments (zero): [Attr(42, 84)] 0x0001 0x002a 0x00000054 0x0000 An example of string and type encoding: [Attr("MyString", typeof(Array), typeof(System.Windows.Forms.Form))] 0x0001 "MyString" "System.Array" "System.Windows.Forms.Form, System.Windows.Forms, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089" 0x0000 As you can see, the full assembly specification of a type is only needed if the type isn't in the current assembly or mscorlib. Note, however, that the C# compiler currently chooses to fully-qualify mscorlib types anyway. An object argument (this binds to the object attribute constructor), and two named arguments (a null string is represented by 0xff and the empty string by 0x00) [Attr((ushort)40, Prop1 = 12, Prop2 = "")] 0x0001 U2 0x0028 0x0002 0x54 I4 "Prop1" 0x0000000c 0x54 0x51 "Prop2" STRING 0x00 Right, more complicated now. A type array as a fixed argument: [Attr(new[] { typeof(string), typeof(object) })] 0x0001 0x00000002 // the number of elements "System.String" "System.Object" 0x0000 An enum value, which is simply represented using the underlying value. The CLR works out that it's an enum using information in the attribute constructor signature: [Attr(MyEnum.Val1)] 0x0001 0x00000001 0x0000 And finally, a null array, and an object array as a named argument: [Attr((Type[])null, ObjectArray = new object[] { (byte)2, typeof(decimal), null, MyEnum.Val2 })] 0x0001 0xffffffff 0x0001 0x53 SZARRAY 0x51 "ObjectArray" 0x00000004 U1 0x02 0x50 "System.Decimal" STRING 0xff 0x55 "MyEnum" 0x00000002 As you'll notice, a null object is encoded as a null string value, and a null array is represented using a length of -1 (0xffffffff). How does this affect C#? So, we can now explain why the limits on attribute arguments are so strict in C#. Attribute specification blobs are limited to basic numbers, enums, types, and arrays. As you can see, this is because the raw CLR encoding can only accommodate those types. Special byte patterns have to be used to indicate object, string, Type, or enum values in named arguments; you can't specify an arbitary object type, as there isn't a generalised way of encoding the resulting value in the specification blob. In particular, decimal values can't be encoded, as it isn't a 'built-in' CLR type that has a native representation (you'll notice that decimal constants in C# programs are compiled as several integer arguments to DecimalConstantAttribute). Jagged arrays also aren't natively supported, although you can get around it by using an array as a value to an object argument: [Attr(new object[] { new object[] { new Type[] { typeof(string) } }, 42 })] Finally... Phew! That was a bit longer than I thought it would be. Custom attribute encodings are complicated! Hopefully this series has been an informative look at what exactly goes on inside a .NET assembly. In the next blog posts, I'll be carrying on with the 'Inside Red Gate' series.

    Read the article

  • Automating deployments with the SQL Compare command line

    - by Jonathan Hickford
    In my previous article, “Five Tips to Get Your Organisation Releasing Software Frequently” I looked at how teams can automate processes to speed up release frequency. In this post, I’m looking specifically at automating deployments using the SQL Compare command line. SQL Compare compares SQL Server schemas and deploys the differences. It works very effectively in scenarios where only one deployment target is required – source and target databases are specified, compared, and a change script is automatically generated and applied. But if multiple targets exist, and pressure to increase the frequency of releases builds, this solution quickly becomes unwieldy.   This is where SQL Compare’s command line comes into its own. I’ve put together a PowerShell script that loops through the Servers table and pulls out the server and database, these are then passed to sqlcompare.exe to be used as target parameters. In the example the source database is a scripts folder, a folder structure of scripted-out database objects used by both SQL Source Control and SQL Compare. The script can easily be adapted to use schema snapshots.     -- Create a DeploymentTargets database and a Servers table CREATE DATABASE DeploymentTargets GO USE DeploymentTargets GO CREATE TABLE [dbo].[Servers]( [id] [int] IDENTITY(1,1) NOT NULL, [serverName] [nvarchar](50) NULL, [environment] [nvarchar](50) NULL, [databaseName] [nvarchar](50) NULL, CONSTRAINT [PK_Servers] PRIMARY KEY CLUSTERED ([id] ASC) ) GO -- Now insert your target server and database details INSERT INTO dbo.Servers ( serverName , environment , databaseName) VALUES ( N'myserverinstance' , N'myenvironment1' , N'mydb1') INSERT INTO dbo.Servers ( serverName , environment , databaseName) VALUES ( N'myserverinstance' , N'myenvironment2' , N'mydb2') Here’s the PowerShell script you can adapt for yourself as well. # We're holding the server names and database names that we want to deploy to in a database table. # We need to connect to that server to read these details $serverName = "" $databaseName = "DeploymentTargets" $authentication = "Integrated Security=SSPI" #$authentication = "User Id=xxx;PWD=xxx" # If you are using database authentication instead of Windows authentication. # Path to the scripts folder we want to deploy to the databases $scriptsPath = "SimpleTalk" # Path to SQLCompare.exe $SQLComparePath = "C:\Program Files (x86)\Red Gate\SQL Compare 10\sqlcompare.exe" # Create SQL connection string, and connection $ServerConnectionString = "Data Source=$serverName;Initial Catalog=$databaseName;$authentication" $ServerConnection = new-object system.data.SqlClient.SqlConnection($ServerConnectionString); # Create a Dataset to hold the DataTable $dataSet = new-object "System.Data.DataSet" "ServerList" # Create a query $query = "SET NOCOUNT ON;" $query += "SELECT serverName, environment, databaseName " $query += "FROM dbo.Servers; " # Create a DataAdapter to populate the DataSet with the results $dataAdapter = new-object "System.Data.SqlClient.SqlDataAdapter" ($query, $ServerConnection) $dataAdapter.Fill($dataSet) | Out-Null # Close the connection $ServerConnection.Close() # Populate the DataTable $dataTable = new-object "System.Data.DataTable" "Servers" $dataTable = $dataSet.Tables[0] #For every row in the DataTable $dataTable | FOREACH-OBJECT { "Server Name: $($_.serverName)" "Database Name: $($_.databaseName)" "Environment: $($_.environment)" # Compare the scripts folder to the database and synchronize the database to match # NB. Have set SQL Compare to abort on medium level warnings. $arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/AbortOnWarnings:Medium") # + @("/sync" ) # Commented out the 'sync' parameter for safety, write-host $arguments & $SQLComparePath $arguments "Exit Code: $LASTEXITCODE" # Some interesting variations # Check that every database matches a folder. # For example this might be a pre-deployment step to validate everything is at the same baseline state. # Or a post deployment script to validate the deployment worked. # An exit code of 0 means the databases are identical. # # $arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/Assertidentical") # Generate a report of the difference between the folder and each database. Generate a SQL update script for each database. # For example use this after the above to generate upgrade scripts for each database # Examine the warnings and the HTML diff report to understand how the script will change objects # #$arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/ScriptFile:update_$($_.environment+"_"+$_.databaseName).sql", "/report:update_$($_.environment+"_"+$_.databaseName).html" , "/reportType:Interactive", "/showWarnings", "/include:Identical") } It’s worth noting that the above example generates the deployment scripts dynamically. This approach should be problem-free for the vast majority of changes, but it is still good practice to review and test a pre-generated deployment script prior to deployment. An alternative approach would be to pre-generate a single deployment script using SQL Compare, and run this en masse to multiple targets programmatically using sqlcmd, or using a tool like SQL Multi Script.  You can use the /ScriptFile, /report, and /showWarnings flags to generate change scripts, difference reports and any warnings.  See the commented out example in the PowerShell: #$arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/ScriptFile:update_$($_.environment+"_"+$_.databaseName).sql", "/report:update_$($_.environment+"_"+$_.databaseName).html" , "/reportType:Interactive", "/showWarnings", "/include:Identical") There is a drawback of running a pre-generated deployment script; it assumes that a given database target hasn’t drifted from its expected state. Often there are (rightly or wrongly) many individuals within an organization who have permissions to alter the production database, and changes can therefore be made outside of the prescribed development processes. The consequence is that at deployment time, the applied script has been validated against a target that no longer represents reality. The solution here would be to add a check for drift prior to running the deployment script. This is achieved by using sqlcompare.exe to compare the target against the expected schema snapshot using the /Assertidentical flag. Should this return any differences (sqlcompare.exe Exit Code 79), a drift report is outputted instead of executing the deployment script.  See the commented out example. # $arguments = @("/scripts1:$($scriptsPath)", "/server2:$($_.serverName)", "/database2:$($_.databaseName)", "/Assertidentical") Any checks and processes that should be undertaken prior to a manual deployment, should also be happen during an automated deployment. You might think about triggering backups prior to deployment – even better, automate the verification of the backup too.   You can use SQL Compare’s command line interface along with PowerShell to automate multiple actions and checks that you need in your deployment process. Automation is a practical solution where multiple targets and a higher release cadence come into play. As we know, with great power comes great responsibility – responsibility to ensure that the necessary checks are made so deployments remain trouble-free.  (The code sample supplied in this post automates the simple dynamic deployment case – if you are considering more advanced automation, e.g. the drift checks, script generation, deploying to large numbers of targets and backup/verification, please email me at [email protected] for further script samples or if you have further questions)

    Read the article

  • Agile Testing Days 2012 – Day 3 – Agile or agile?

    - by Chris George
    Another early start for my last Lean Coffee of the conference, and again it was not wasted. We had some really interesting discussions around how to determine what test automation is useful, if agile is not faster, why do it? and a rather existential discussion on whether unicorns exist! First keynote of the day was entitled “Fast Feedback Teams” by Ola Ellnestam. Again this relates nicely to the releasing faster talk on day 2, and something that we are looking at and some teams are actively trying. Introducing the notion of feedback, Ola describes a game he wrote for his eldest child. It was a simple game where every time he clicked a button, it displayed “You’ve Won!”. He then changed it to be a Win-Lose-Win-Lose pattern and watched the feedback from his son who then twigged the pattern and got his younger brother to play, alternating turns… genius! (must do that with my children). The idea behind this was that you need that feedback loop to learn and progress. If you are not getting the feedback you need to close that loop. An interesting point Ola made was to solve problems BEFORE writing software. It may be that you don’t have to write anything at all, perhaps it’s a communication/training issue? Perhaps the problem can be solved another way. Writing software, although it’s the business we are in, is expensive, and this should be taken into account. He again mentions frequent releases, and how they should be made as soon as stuff is ready to be released, don’t leave stuff on the shelf cause it’s not earning you anything, money or data. I totally agree with this and it’s something that we will be aiming for moving forwards. “Exceptions, Assumptions and Ambiguity: Finding the truth behind the story” by David Evans started off very promising by making references to ‘Grim up North’ referring to the north of England. Not sure it was appreciated by most of the audience, but it made me laugh! David explained how there are always risks associated with exceptions, giving the example of a one-way road near where he lives, with an exception sign giving rights to coaches to go the wrong way. Therefore you could merrily swing around the corner of the one way road straight into a coach! David showed the danger in making assumptions with lyrical quotes from Lola by The Kinks “I’m glad I’m a man, and so is Lola” and with a picture of a toilet flush that needed instructions to operate the full and half flush. With this particular flush, you pulled the handle all the way down to half flush, and half way down to full flush! hmmm, a bit of a crappy user experience methinks! Then through a clever use of a passage from the Jabberwocky, David then went onto show how mis-translation/ambiguity is the can completely distort the original meaning of something, and this is a real enemy of software development. This was all helping to demonstrate that the term Story is often heavily overloaded in the Agile world, and should really be stripped back to what it is really for, stating a business problem, and offering a technical solution. Therefore a story could be worded as “In order to {make some improvement}, we will { do something}”. The first ‘in order to’ statement is stakeholder neutral, and states the problem through requesting an improvement to the software/process etc. The second part of the story is the verb, the doing bit. So to achieve the ‘improvement’ which is not currently true, we will do something to make this true in the future. My PM is very interested in this, and he’s observed some of the problems of overloading stories so I’m hoping between us we can use some of David’s suggestions to help clarify our stories better. The second keynote of the day (and our last) proved to be the most entertaining and exhausting of the conference for me. “The ongoing evolution of testing in agile development” by Scott Barber. I’ve never had the pleasure of seeing Scott before… OMG I would love to have even half of the energy he has! What struck me during this presentation was Scott’s explanation of how testing has become the role/job that it is (largely) today, and how this has led to the need for ‘methodologies’ to make dev and test work! The argument that we should be trying to converge the roles again is a very valid one, and one that a couple of the teams at work are actively doing with great results. Making developers as responsible for quality as testers is something that has been lost over the years, but something that we are now striving to achieve. The idea that we (testers) should be testing experts/specialists, not testing ‘union members’, supports this idea so the entire team works on all aspects of a feature/product, with the ‘specialists’ taking the lead and advising/coaching the others. This leads to better propagation of information around the team, a greater holistic understanding of the project and it allows the team to continue functioning if some of it’s members are off sick, for example. Feeling somewhat drained from Scott’s keynote (but at the same time excited that alot of the points he raised supported actions we are taking at work), I headed into my last presentation for Agile Testing Days 2012 before having to make my way to Tegel to catch the flight home. “Thinking and working agile in an unbending world” with Pete Walen was a talk I was not going to miss! Having spoken to Pete several times during the past few days, I was looking forward to hearing what he was going to say, and I was not disappointed. Pete started off by trying to separate the definitions of ‘Agile’ as in the methodology, and ‘agile’ as in the adjective by pronouncing them the ‘english’ and ‘american’ ways. So Agile pronounced (Ajyle) and agile pronounced (ajul). There was much confusion around what the hell he was talking about, although I thought it was quite clear. Agile – Software development methodology agile – Marked by ready ability to move with quick easy grace; Having a quick resourceful and adaptable character. Anyway, that aside (although it provided a few laughs during the presentation), the point was that many teams that claim to be ‘Agile’ but are not, in fact, ‘agile’ by nature. Implementing ‘Agile’ methodologies that are so prescriptive actually goes against the very nature of Agile development where a team should anticipate, adapt and explore. Pete made a valid point that very few companies intentionally put up roadblocks to impede work, so if work is being blocked/delayed, why? This is where being agile as a team pays off because the team can inspect what’s going on, explore options and adapt their processes. It is through experimentation (and that means trying and failing as well as trying and succeeding) that a team will improve and grow leading to focussing on what really needs to be done to achieve X. So, that was it, the last talk of our conference. I was gutted that we had to miss the closing keynote from Matt Heusser, as Matt was another person I had spoken too a few times during the conference, but the flight would not wait, and just as well we left when we did because the traffic was a nightmare! My Takeaway Triple from Day 3: Release often and release small – don’t leave stuff on the shelf Keep the meaning of the word ‘agile’ in mind when working in ‘Agile Look at testing as more of a skill than a role  

    Read the article

  • Learnings from trying to write better software: Loud errors from the very start

    - by theo.spears
    Microsoft made a very small number of backwards incompatible changes between .NET 1.1 and 2.0, because they wanted to make it as easy and safe as possible to port applications to the new runtime. (Here’s a list.) However, one thing they did change was what happens when a background thread fails with an unhanded exception - in .NET 1.1 nothing happened, the thread terminated, and the application continued oblivious. Try the same trick in .NET 2.0 and the entire application, including all threads, will rudely terminate. There are three reasons for this. Firstly if a background thread has crashed, it may have left the entire application in an inconsistent state, in a way that will affect other threads. It’s better to terminate the entire application than continue and have the application perform actions based on a broken state, for example take customer orders, or write corrupt files to disk.  Secondly, during software development, it is far better for errors to be loud and obtrusive. Even if you have unit tests and integration tests (and you should), a key part of ensuring software works properly is to actually try using it, both through systematic testing and through the casual use all software gets by its developers during use. Subtle errors are easy to miss if you are not actually doing real work using the application, loud errors are obvious. Thirdly, and most importantly, even if catching and swallowing exceptions indiscriminately doesn't cause any problems in your application, the presence of unexpected exceptions shows you do not fully understand the behavior of your code. The currently released version of your application may be absolutely correct. However, because your mental model of the behavior is wrong, any future change you make to the program could and probably will introduce critical errors.  This applies to more than just exceptions causing threads to exit, any unexpected state should make the application blow up in an un-ignorable way. The worst thing you can do is silently swallow errors and continue. And let's be clear, writing to a log file does not count as blowing up in an un-ignorable way.  This is all simple as long as the call stack only contains your code, but when your functions start to be called by third party or .NET framework code, it's surprisingly easy for exceptions to start vanishing. Let's look at two examples.   1. Windows forms drag drop events  Usually if you throw an exception from a winforms event handler it will bring up the "application has crashed" dialog with abort and continue options. This is a good default behavior - the error is big and loud, but it is possible for the user to ignore the error and hopefully save their data, if somehow this bug makes it past testing. However drag and drop are different - throw an exception from one of these and it will just be silently swallowed with no explanation.  By the way, it's not just drag and drop events. Timer events do it too.  You can research how exceptions are treated in different handlers and code appropriately, but the safest and most user friendly approach is to always catch exceptions in your event handlers and show your own error message. I'll talk about one good approach to handling these exceptions at the end of this post.   2. SSMS integration for SQL Tab Magic  A while back wrote an SSMS add-in called SQL Tab Magic (learn more about the process here). It works by listening to certain SSMS events and remembering what documents are opened and closed. I deployed it internally and it was used for a few months by a number of people without problems, so I was reasonably confident in its quality. Before releasing I made a few cleanups, including introducing error reporting. Bam. A few days later I was looking at over 1,000 error reports in my inbox. In turns out I wasn't handling table designers properly. The exceptions were there, but again SSMS was helpfully swallowing them all for me, so I was blissfully unaware. Had I made my errors loud from the start, I would have noticed these issues long before and fixed them.   Handling exceptions  Now you are systematically catching exceptions throughout your application, you need to do something with them. I've tried 3 options: log them, alert the user, and automatically send them home.  There are a few good options for logging in .NET. The most widespread is Apache log4net, which provides a very capable and configurable logging framework. There is also NLog which has a compatible interface, with a greater emphasis on fluent rather than XML configuration.  Alerting the user serves two purposes. Firstly it means they understand their action has failed to they don't just assume it worked (Silent file copy failure is a problem if you then delete the originals) or that they should keep waiting for a background task to complete. Secondly, it means the users can report the bug to your support team, and then you can fix it. This means the message you show the user should contain the information you need as a developer to identify and fix it. And the user will probably just send you a screenshot of the dialog, so it shouldn't be hidden by scroll bars.  This leads us to the third option, automatically sending error reports home. By automatic I mean with minimal effort on the part of the user, rather than doing it silently behind their backs. The advantage of this is you can send back far more detailed and precise information than you can expect a user to include in an email, and by making it easier to report errors, you make it more likely users will do so.  We do this using a great tool called SmartAssembly (full disclosure: this is a product made by Red Gate). It captures complete stack traces including the values of all local variables and then allows the user to send all this information back with a single click. We also capture log files to help understand what lead up to the error. We then use the free SmartAssembly Sync for Jira to dedupe these reports and raise them as bugs in our bug tracking system.  The combined effect of loud errors during development and then automatic error reporting once software is deployed allows us to find and fix more bugs, correct misunderstandings on how our software works, and overall is a key piece in delivering higher quality software. However it is no substitute for having motivated cunning testers in the building - and we're looking to hire more of those too.   If you found this post interesting you should follow me on twitter.  

    Read the article

  • We've completed the first iteration

    - by CliveT
    There are a lot of features in C# that are implemented by the compiler and not by the underlying platform. One such feature is a lambda expression. Since local variables cannot be accessed once the current method activation finishes, the compiler has to go out of its way to generate a new class which acts as a home for any variable whose lifetime needs to be extended past the activation of the procedure. Take the following example:     Random generator = new Random();     Func func = () = generator.Next(10); In this case, the compiler generates a new class called c_DisplayClass1 which is marked with the CompilerGenerated attribute. [CompilerGenerated] private sealed class c__DisplayClass1 {     // Fields     public Random generator;     // Methods     public int b__0()     {         return this.generator.Next(10);     } } Two quick comments on this: (i)    A display was the means that compilers for languages like Algol recorded the various lexical contours of the nested procedure activations on the stack. I imagine that this is what has led to the name. (ii)    It is a shame that the same attribute is used to mark all compiler generated classes as it makes it hard to figure out what they are being used for. Indeed, you could imagine optimisations that the runtime could perform if it knew that classes corresponded to certain high level concepts. We can see that the local variable generator has been turned into a field in the class, and the body of the lambda expression has been turned into a method of the new class. The code that builds the Func object simply constructs an instance of this class and initialises the fields to their initial values.     c__DisplayClass1 class2 = new c__DisplayClass1();     class2.generator = new Random();     Func func = new Func(class2.b__0); Reflector already contains code to spot this pattern of code and reproduce the form containing the lambda expression, so this is example is correctly decompiled. The use of compiler generated code is even more spectacular in the case of iterators. C# introduced the idea of a method that could automatically store its state between calls, so that it can pick up where it left off. The code can express the logical flow with yield return and yield break denoting places where the method should return a particular value and be prepared to resume.         {             yield return 1;             yield return 2;             yield return 3;         } Of course, there was already a .NET pattern for expressing the idea of returning a sequence of values with the computation proceeding lazily (in the sense that the work for the next value is executed on demand). This is expressed by the IEnumerable interface with its Current property for fetching the current value and the MoveNext method for forcing the computation of the next value. The sequence is terminated when this method returns false. The C# compiler links these two ideas together so that an IEnumerator returning method using the yield keyword causes the compiler to produce the implementation of an Iterator. Take the following piece of code.         IEnumerable GetItems()         {             yield return 1;             yield return 2;             yield return 3;         } The compiler implements this by defining a new class that implements a state machine. This has an integer state that records which yield point we should go to if we are resumed. It also has a field that records the Current value of the enumerator and a field for recording the thread. This latter value is used for optimising the creation of iterator instances. [CompilerGenerated] private sealed class d__0 : IEnumerable, IEnumerable, IEnumerator, IEnumerator, IDisposable {     // Fields     private int 1__state;     private int 2__current;     public Program 4__this;     private int l__initialThreadId; The body gets converted into the code to construct and initialize this new class. private IEnumerable GetItems() {     d__0 d__ = new d__0(-2);     d__.4__this = this;     return d__; } When the class is constructed we set the state, which was passed through as -2 and the current thread. public d__0(int 1__state) {     this.1__state = 1__state;     this.l__initialThreadId = Thread.CurrentThread.ManagedThreadId; } The state needs to be set to 0 to represent a valid enumerator and this is done in the GetEnumerator method which optimises for the usual case where the returned enumerator is only used once. IEnumerator IEnumerable.GetEnumerator() {     if ((Thread.CurrentThread.ManagedThreadId == this.l__initialThreadId)               && (this.1__state == -2))     {         this.1__state = 0;         return this;     } The state machine itself is implemented inside the MoveNext method. private bool MoveNext() {     switch (this.1__state)     {         case 0:             this.1__state = -1;             this.2__current = 1;             this.1__state = 1;             return true;         case 1:             this.1__state = -1;             this.2__current = 2;             this.1__state = 2;             return true;         case 2:             this.1__state = -1;             this.2__current = 3;             this.1__state = 3;             return true;         case 3:             this.1__state = -1;             break;     }     return false; } At each stage, the current value of the state is used to determine how far we got, and then we generate the next value which we return after recording the next state. Finally we return false from the MoveNext to signify the end of the sequence. Of course, that example was really simple. The original method body didn't have any local variables. Any local variables need to live between the calls to MoveNext and so they need to be transformed into fields in much the same way that we did in the case of the lambda expression. More complicated MoveNext methods are required to deal with resources that need to be disposed when the iterator finishes, and sometimes the compiler uses a temporary variable to hold the return value. Why all of this explanation? We've implemented the de-compilation of iterators in the current EAP version of Reflector (7). This contrasts with previous version where all you could do was look at the MoveNext method and try to figure out the control flow. There's a fair amount of things we have to do. We have to spot the use of a CompilerGenerated class which implements the Enumerator pattern. We need to go to the class and figure out the fields corresponding to the local variables. We then need to go to the MoveNext method and try to break it into the various possible states and spot the state transitions. We can then take these pieces and put them back together into an object model that uses yield return to show the transition points. After that Reflector can carry on optimising using its usual optimisations. The pattern matching is currently a little too sensitive to changes in the code generation, and we only do a limited analysis of the MoveNext method to determine use of the compiler generated fields. In some ways, it is a pity that iterators are compiled away and there is no metadata that reflects the original intent. Without it, we are always going to dependent on our knowledge of the compiler's implementation. For example, we have noticed that the Async CTP changes the way that iterators are code generated, so we'll have to do some more work to support that. However, with that warning in place, we seem to do a reasonable job of decompiling the iterators that are built into the framework. Hopefully, the EAP will give us a chance to find examples where we don't spot the pattern correctly or regenerate the wrong code, and we can improve things. Please give it a go, and report any problems.

    Read the article

  • MySQL MyISAM table performance... painfully, painfully slow

    - by Salman A
    I've got a table structure that can be summarized as follows: pagegroup * pagegroupid * name has 3600 rows page * pageid * pagegroupid * data references pagegroup; has 10000 rows; can have anything between 1-700 rows per pagegroup; the data column is of type mediumtext and the column contains 100k - 200kbytes data per row userdata * userdataid * pageid * column1 * column2 * column9 references page; has about 300,000 rows; can have about 1-50 rows per page The above structure is pretty straight forwad, the problem is that that a join from userdata to page group is terribly, terribly slow even though I have indexed all columns that should be indexed. The time needed to run a query for such a join (userdata inner_join page inner_join pagegroup) exceeds 3 minutes. This is terribly slow considering the fact that I am not selecting the data column at all. Example of the query that takes too long: SELECT userdata.column1, pagegroup.name FROM userdata INNER JOIN page USING( pageid ) INNER JOIN pagegroup USING( pagegroupid ) Please help by explaining why does it take so long and what can i do to make it faster. Edit #1 Explain returns following gibberish: id select_type table type possible_keys key key_len ref rows Extra 1 SIMPLE userdata ALL pageid 372420 1 SIMPLE page eq_ref PRIMARY,pagegroupid PRIMARY 4 topsecret.userdata.pageid 1 1 SIMPLE pagegroup eq_ref PRIMARY PRIMARY 4 topsecret.page.pagegroupid 1 Edit #2 SELECT u.field2, p.pageid FROM userdata u INNER JOIN page p ON u.pageid = p.pageid; /* 0.07 sec execution, 6.05 sec fecth */ id select_type table type possible_keys key key_len ref rows Extra 1 SIMPLE u ALL pageid 372420 1 SIMPLE p eq_ref PRIMARY PRIMARY 4 topsecret.u.pageid 1 Using index SELECT p.pageid, g.pagegroupid FROM page p INNER JOIN pagegroup g ON p.pagegroupid = g.pagegroupid; /* 9.37 sec execution, 60.0 sec fetch */ id select_type table type possible_keys key key_len ref rows Extra 1 SIMPLE g index PRIMARY PRIMARY 4 3646 Using index 1 SIMPLE p ref pagegroupid pagegroupid 5 topsecret.g.pagegroupid 3 Using where Moral of the story Keep medium/long text columns in a separate table if you run into performance problems such as this one.

    Read the article

  • Ada and 'The Book'

    - by Phil Factor
    The long friendship between Charles Babbage and Ada Lovelace created one of the most exciting and mysterious of collaborations ever to have resulted in a technological breakthrough. The fireworks that created by the collision of two prodigious mathematical and creative talents resulted in an invention, the Analytical Engine, which went on to change society fundamentally. However, beyond that, we just don't know what the bulk of their collaborative work was about:;  it was done in strictest secrecy. Even the known outcome of their friendship, the first programmable computer, was shrouded in mystery. At the time, nobody, except close friends and family, had any idea of Ada Byron's contribution to the invention of the ‘Engine’, and how to program it. Her great insight was published in August 1843, under the initials AAL, standing for Ada Augusta Lovelace, her title then being the Countess of Lovelace. It was contained in a lengthy ‘note’ to her translation of a publication that remains the best description of Babbage's amazing Analytical Engine. The secret identity of the person behind those enigmatic initials was finally revealed by Prince de Polignac who, seventy years later, wrote to Ada's daughter to seek confirmation that her mother had, indeed, been the author of the brilliant sentences that described so accurately how Babbage's mechanical computer could be programmed with punch-cards. L.F. Menabrea's paper on the Analytical Engine first appeared in the 'Bibliotheque Universelle de Geneve' in October 1842, and Ada translated it anonymously for Taylor's 'Scientific Memoirs'. Charles Babbage was surprised that she had not written an original paper as she already knew a surprising amount about the way the machine worked. He persuaded her to at least write some explanatory notes. These notes ended up extending to four times the length of the original article and represented the first published account of how a machine could be programmed to perform any calculation. Her example of programming the Bernoulli sequence would have worked on the Analytical engine had the device’s construction been completed, and gave Ada an unassailable claim to have invented the art of programming. What was the reason for Ada's secrecy? She was the only legitimate child of Lord Byron, who was probably the best known celebrity of the age, so she was already famous. She was a senior aristocrat, with titles, a fortune in money and vast estates in the Midlands. She had political influence, and was the cousin of Lord Melbourne, who was the Prime Minister at that time. She was friendly with the young Queen Victoria. Her mathematical activities were a pastime, and not one that would be considered by others to be in keeping with her roles and responsibilities. You wouldn't dare to dream up a fictional heroine like Ada. She was dazzlingly beautiful and talented. She could speak several languages fluently, and play some musical instruments with professional skill. Contemporary accounts refer to her being 'accomplished in science, art and literature'. On top of that, she was a brilliant mathematician, a talent inherited from her mother, Annabella Milbanke. In her mother's circle of literary and scientific friends was Charles Babbage, and Ada's friendship with him dates from her teenage zest for Mathematics. She was one of the first people he'd ever met who understood what he had attempted to achieve with the 'Difference Engine', and with whom he could converse as intellectual equals. He arranged for her to have an education from the most talented academics in the country. Ada melted the heart of the cantankerous genius to the point that he became a faithful and loyal father-figure to her. She was one of the very few who could grasp the principles of the later, and very different, ‘Analytical Engine’ which was designed from the start to tackle a variety of tasks. Sadly, Ada Byron's life ended less than a decade after completing the work that assured her long-term fame, in November 1852. She was dying of cancer, her gambling habits had caused her to run up huge debts, she'd had more than one affairs, and she was being blackmailed. Her brilliant but unempathic mother was nursing her in her final illness, destroying her personal letters and records, and repaying her debts. Her husband was distraught but helpless. Charles Babbage, however, maintained his steadfast paternalistic friendship to the end. She appointed her loyal friend to be her executor. For years, she and Babbage had been working together on a secret project, known only as 'The Book'. We have a clue to what it was in a letter written by her nine years earlier, on 11th August 1843. It was a joint project by herself and Lord Lovelace, her husband, and was intended to involve Babbage's 'undivided energies'. It involved 'consulting your Engine' (it required Babbage’s computer). The letter gives no hint about the project except for the high-minded nature of its purpose, and its highly mathematical nature.  From then on, the surviving correspondence between the two gives only veiled references to 'The Book'. There isn't much, since Babbage later destroyed any letters that could have damaged her reputation within the Establishment. 'I cannot spare the book today, which I am very sorry for. At the moment I want it for constant reference, but I think you can have it tomorrow' (Oct 1844)  And 'I will send you the book directly, and you can say, when you receive it, how long you will want to keep it'. (Nov 1844)  The two of them were obviously intent on the work: She writes, four years later, 'I have an engagement for Wednesday which will prevent me from attending to your wishes about the book' (Dec 1848). This was something that they both needed to work on, but could not do in parallel: 'I will send the book on Tuesday, and it can be left with you till Friday' (11 Feb 1849). After six years work, it had been so well-handled that it was beginning to fall apart: 'Don't forget the new cover you promised for the book. The poor book is very shabby and wants one' (20 Sept 1849). So what was going on? The word 'book' was not a code-word: it was a real book, probably a 'printer's blank', plain paper, but properly bound so printers and publishers could show off how the published work might look. The hints from the correspondence are of advanced mathematics. It is obvious that the book was travelling between them, back and forth, each one working on it for less than a week before passing it back. Ada and her husband were certainly involved in gambling large sums of money on the horses, and so most biographers have concluded that the three of them were trying to calculate the mathematical odds on the horses. This theory has three large problems. Firstly, Ada's original letter proposing the project refers to its high-minded nature. Babbage was temperamentally opposed to gambling and would scarcely have given so much time to the project, even though he was devoted to Ada. Secondly, Babbage would have very soon have realized the hopelessness of trying to beat the bookies. This sort of betting never attracts his type of intellectual background. The third problem is that any work on calculating the odds on horses would not need a well-thumbed book to pass back and forth between them; they would have not had to work in series. The original project was instigated by Ada, along with her husband, William King-Noel, 1st Earl of Lovelace. Charles Babbage was invited to join the project after the couple had come up with the idea. What could William have contributed? One might assume that William was a Bertie Wooster character, addicted only to the joys of the turf, but this was far from the truth. He was a scientist, a Cambridge graduate who was later elected to be a Fellow of the Royal Society. After Eton, he went to Trinity College, Cambridge. On graduation, he entered the diplomatic service and acted as secretary under Lord Nugent, who was Lord Commissioner of the Ionian Islands. William was very friendly with Babbage too, able to discuss scientific matters on equal terms. He was a capable engineer who invented a process for bending large timbers by the application of steam heat. He delivered a paper to the Institution of Civil Engineers in 1849, and received praise from the great engineer, Isambard Kingdom Brunel. As well as being Lord Lieutenant of the County of Surrey for most of Victoria's reign, he had time for a string of scientific and engineering achievements. Whatever the project was, it is unlikely that William was a junior partner. After Ada's death, the project disappeared. Then, two years later, Babbage, through one of his occasional outbursts of temper, demonstrated that he was able to decrypt one of the most powerful of secret codes, Vigenère's autokey cipher.  All contemporary diplomatic and military messages used a variant of this cipher. Babbage had made three important discoveries, namely, the mathematical law of this cipher, the principle of the key periodicity, and the technique of the symmetry of position. The technique is now known as the Kasiski examination, also called the Kasiski test, but Babbage got there first. At one time, he listed amongst his future projects, the writing of a book 'The Philosophy of Decyphering', but it never came to anything. This discovery was going to change the course of history, since it was used to decipher the Russians’ military dispatches in the Crimean war. Babbage himself played a role during the Crimean War as a cryptographical adviser to his friend, Rear-Admiral Sir Francis Beaufort of the Admiralty. This is as much as we can be certain about in trying to make sense of the bulk of the time that Charles Babbage and Ada Lovelace worked together. Nine years of intensive work, involving the 'Engine' and a great deal of mathematics and research seems to have been lost: or has it? I've argued in the past http://www.simple-talk.com/community/blogs/philfactor/archive/2008/06/13/59614.aspx that the cracking of the Vigenère autokey cipher, was a fundamental motive behind the British Government's support and funding of the 'Difference Engine'. The Duke of Wellington, whose understanding of the military significance of being able to read enemy dispatches, was the most steadfast advocate of the project. If the three friends were actually doing the work of cracking codes by mathematical techniques that used the techniques of key periodicity, and symmetry of position (the use of a book being passed quickly to and fro is very suggestive), intending to then use the 'Engine' to do the routine cracking of each dispatch, then this is a rather different story. The project was Ada and William's idea. (William had served in the diplomatic service and would be familiar with the use of codes). This makes Ada Lovelace the initiator of a project which, by giving both Britain, and probably the USA, a diplomatic and military advantage in the second part of the Nineteenth century, changed world history. Ada would never have wanted any credit for cracking the cipher, and developing the method that rendered all contemporary military and diplomatic ciphering techniques nugatory; quite the reverse. And it is clear from the gaps in the record of the letters between the collaborators that the evidence was destroyed, probably on her request by her irascible but intensely honorable executor, Charles Babbage. Charles Babbage toyed with the idea of going public, but the Crimean war put an end to that. The British Government had a valuable secret, and intended to keep it that way. Ada and Charles had quite often discussed possible moneymaking projects that would fund the development of the Analytic Engine, the first programmable computer, but their secret work was never in the running as a potential cash cow. I suspect that the British Government was, even then, working on the concealment of a discovery whose value to the nation depended on it remaining so. The success of code-breaking in the Crimean war, and the American Civil war, led to the British and Americans  subsequently giving much more weight and funding to the science of decryption. Paradoxically, this makes Ada's contribution even closer to the creation of Colossus, the first digital computer, at Bletchley Park, specifically to crack the Nazi’s secret codes.

    Read the article

  • Inside the Concurrent Collections: ConcurrentBag

    - by Simon Cooper
    Unlike the other concurrent collections, ConcurrentBag does not really have a non-concurrent analogy. As stated in the MSDN documentation, ConcurrentBag is optimised for the situation where the same thread is both producing and consuming items from the collection. We'll see how this is the case as we take a closer look. Again, I recommend you have ConcurrentBag open in a decompiler for reference. Thread Statics ConcurrentBag makes heavy use of thread statics - static variables marked with ThreadStaticAttribute. This is a special attribute that instructs the CLR to scope any values assigned to or read from the variable to the executing thread, not globally within the AppDomain. This means that if two different threads assign two different values to the same thread static variable, one value will not overwrite the other, and each thread will see the value they assigned to the variable, separately to any other thread. This is a very useful function that allows for ConcurrentBag's concurrency properties. You can think of a thread static variable: [ThreadStatic] private static int m_Value; as doing the same as: private static Dictionary<Thread, int> m_Values; where the executing thread's identity is used to automatically set and retrieve the corresponding value in the dictionary. In .NET 4, this usage of ThreadStaticAttribute is encapsulated in the ThreadLocal class. Lists of lists ConcurrentBag, at its core, operates as a linked list of linked lists: Each outer list node is an instance of ThreadLocalList, and each inner list node is an instance of Node. Each outer ThreadLocalList is owned by a particular thread, accessible through the thread local m_locals variable: private ThreadLocal<ThreadLocalList<T>> m_locals It is important to note that, although the m_locals variable is thread-local, that only applies to accesses through that variable. The objects referenced by the thread (each instance of the ThreadLocalList object) are normal heap objects that are not specific to any thread. Thinking back to the Dictionary analogy above, if each value stored in the dictionary could be accessed by other means, then any thread could access the value belonging to other threads using that mechanism. Only reads and writes to the variable defined as thread-local are re-routed by the CLR according to the executing thread's identity. So, although m_locals is defined as thread-local, the m_headList, m_nextList and m_tailList variables aren't. This means that any thread can access all the thread local lists in the collection by doing a linear search through the outer linked list defined by these variables. Adding items So, onto the collection operations. First, adding items. This one's pretty simple. If the current thread doesn't already own an instance of ThreadLocalList, then one is created (or, if there are lists owned by threads that have stopped, it takes control of one of those). Then the item is added to the head of that thread's list. That's it. Don't worry, it'll get more complicated when we account for the other operations on the list! Taking & Peeking items This is where it gets tricky. If the current thread's list has items in it, then it peeks or removes the head item (not the tail item) from the local list and returns that. However, if the local list is empty, it has to go and steal another item from another list, belonging to a different thread. It iterates through all the thread local lists in the collection using the m_headList and m_nextList variables until it finds one that has items in it, and it steals one item from that list. Up to this point, the two threads had been operating completely independently. To steal an item from another thread's list, the stealing thread has to do it in such a way as to not step on the owning thread's toes. Recall how adding and removing items both operate on the head of the thread's linked list? That gives us an easy way out - a thread trying to steal items from another thread can pop in round the back of another thread's list using the m_tail variable, and steal an item from the back without the owning thread knowing anything about it. The owning thread can carry on completely independently, unaware that one of its items has been nicked. However, this only works when there are at least 3 items in the list, as that guarantees there will be at least one node between the owning thread performing operations on the list head and the thread stealing items from the tail - there's no chance of the two threads operating on the same node at the same time and causing a race condition. If there's less than three items in the list, then there does need to be some synchronization between the two threads. In this case, the lock on the ThreadLocalList object is used to mediate access to a thread's list when there's the possibility of contention. Thread synchronization In ConcurrentBag, this is done using several mechanisms: Operations performed by the owner thread only take out the lock when there are less than three items in the collection. With three or greater items, there won't be any conflict with a stealing thread operating on the tail of the list. If a lock isn't taken out, the owning thread sets the list's m_currentOp variable to a non-zero value for the duration of the operation. This indicates to all other threads that there is a non-locked operation currently occuring on that list. The stealing thread always takes out the lock, to prevent two threads trying to steal from the same list at the same time. After taking out the lock, the stealing thread spinwaits until m_currentOp has been set to zero before actually performing the steal. This ensures there won't be a conflict with the owning thread when the number of items in the list is on the 2-3 item borderline. If any add or remove operations are started in the meantime, and the list is below 3 items, those operations try to take out the list's lock and are blocked until the stealing thread has finished. This allows a thread to steal an item from another thread's list without corrupting it. What about synchronization in the collection as a whole? Collection synchronization Any thread that operates on the collection's global structure (accessing anything outside the thread local lists) has to take out the collection's global lock - m_globalListsLock. This single lock is sufficient when adding a new thread local list, as the items inside each thread's list are unaffected. However, what about operations (such as Count or ToArray) that need to access every item in the collection? In order to ensure a consistent view, all operations on the collection are stopped while the count or ToArray is performed. This is done by freezing the bag at the start, performing the global operation, and unfreezing at the end: The global lock is taken out, to prevent structural alterations to the collection. m_needSync is set to true. This notifies all the threads that they need to take out their list's lock irregardless of what operation they're doing. All the list locks are taken out in order. This blocks all locking operations on the lists. The freezing thread waits for all current lockless operations to finish by spinwaiting on each m_currentOp field. The global operation can then be performed while the bag is frozen, but no other operations can take place at the same time, as all other threads are blocked on a list's lock. Then, once the global operation has finished, the locks are released, m_needSync is unset, and normal concurrent operation resumes. Concurrent principles That's the essence of how ConcurrentBag operates. Each thread operates independently on its own local list, except when they have to steal items from another list. When stealing, only the stealing thread is forced to take out the lock; the owning thread only has to when there is the possibility of contention. And a global lock controls accesses to the structure of the collection outside the thread lists. Operations affecting the entire collection take out all locks in the collection to freeze the contents at a single point in time. So, what principles can we extract here? Threads operate independently Thread-static variables and ThreadLocal makes this easy. Threads operate entirely concurrently on their own structures; only when they need to grab data from another thread is there any thread contention. Minimised lock-taking Even when two threads need to operate on the same data structures (one thread stealing from another), they do so in such a way such that the probability of actually blocking on a lock is minimised; the owning thread always operates on the head of the list, and the stealing thread always operates on the tail. Management of lockless operations Any operations that don't take out a lock still have a 'hook' to force them to lock when necessary. This allows all operations on the collection to be stopped temporarily while a global snapshot is taken. Hopefully, such operations will be short-lived and infrequent. That's all the concurrent collections covered. I hope you've found it as informative and interesting as I have. Next, I'll be taking a closer look at ThreadLocal, which I came across while analyzing ConcurrentBag. As you'll see, the operation of this class deserves a much closer look.

    Read the article

  • Operator of the week - Assert

    - by Fabiano Amorim
    Well my friends, I was wondering how to help you in a practical way to understand execution plans. So I think I'll talk about the Showplan Operators. Showplan Operators are used by the Query Optimizer (QO) to build the query plan in order to perform a specified operation. A query plan will consist of many physical operators. The Query Optimizer uses a simple language that represents each physical operation by an operator, and each operator is represented in the graphical execution plan by an icon. I'll try to talk about one operator every week, but so as to avoid having to continue to write about these operators for years, I'll mention only of those that are more common: The first being the Assert. The Assert is used to verify a certain condition, it validates a Constraint on every row to ensure that the condition was met. If, for example, our DDL includes a check constraint which specifies only two valid values for a column, the Assert will, for every row, validate the value passed to the column to ensure that input is consistent with the check constraint. Assert  and Check Constraints: Let's see where the SQL Server uses that information in practice. Take the following T-SQL: IF OBJECT_ID('Tab1') IS NOT NULL   DROP TABLE Tab1 GO CREATE TABLE Tab1(ID Integer, Gender CHAR(1))  GO  ALTER TABLE TAB1 ADD CONSTRAINT ck_Gender_M_F CHECK(Gender IN('M','F'))  GO INSERT INTO Tab1(ID, Gender) VALUES(1,'X') GO To the command above the SQL Server has generated the following execution plan: As we can see, the execution plan uses the Assert operator to check that the inserted value doesn't violate the Check Constraint. In this specific case, the Assert applies the rule, 'if the value is different to "F" and different to "M" than return 0 otherwise returns NULL'. The Assert operator is programmed to show an error if the returned value is not NULL; in other words, the returned value is not a "M" or "F". Assert checking Foreign Keys Now let's take a look at an example where the Assert is used to validate a foreign key constraint. Suppose we have this  query: ALTER TABLE Tab1 ADD ID_Genders INT GO  IF OBJECT_ID('Tab2') IS NOT NULL   DROP TABLE Tab2 GO CREATE TABLE Tab2(ID Integer PRIMARY KEY, Gender CHAR(1))  GO  INSERT INTO Tab2(ID, Gender) VALUES(1, 'F') INSERT INTO Tab2(ID, Gender) VALUES(2, 'M') INSERT INTO Tab2(ID, Gender) VALUES(3, 'N') GO  ALTER TABLE Tab1 ADD CONSTRAINT fk_Tab2 FOREIGN KEY (ID_Genders) REFERENCES Tab2(ID) GO  INSERT INTO Tab1(ID, ID_Genders, Gender) VALUES(1, 4, 'X') Let's look at the text execution plan to see what these Assert operators were doing. To see the text execution plan just execute SET SHOWPLAN_TEXT ON before run the insert command. |--Assert(WHERE:(CASE WHEN NOT [Pass1008] AND [Expr1007] IS NULL THEN (0) ELSE NULL END))      |--Nested Loops(Left Semi Join, PASSTHRU:([Tab1].[ID_Genders] IS NULL), OUTER REFERENCES:([Tab1].[ID_Genders]), DEFINE:([Expr1007] = [PROBE VALUE]))           |--Assert(WHERE:(CASE WHEN [Tab1].[Gender]<>'F' AND [Tab1].[Gender]<>'M' THEN (0) ELSE NULL END))           |    |--Clustered Index Insert(OBJECT:([Tab1].[PK]), SET:([Tab1].[ID] = RaiseIfNullInsert([@1]),[Tab1].[ID_Genders] = [@2],[Tab1].[Gender] = [Expr1003]), DEFINE:([Expr1003]=CONVERT_IMPLICIT(char(1),[@3],0)))           |--Clustered Index Seek(OBJECT:([Tab2].[PK]), SEEK:([Tab2].[ID]=[Tab1].[ID_Genders]) ORDERED FORWARD) Here we can see the Assert operator twice, first (looking down to up in the text plan and the right to left in the graphical plan) validating the Check Constraint. The same concept showed above is used, if the exit value is "0" than keep running the query, but if NULL is returned shows an exception. The second Assert is validating the result of the Tab1 and Tab2 join. It is interesting to see the "[Expr1007] IS NULL". To understand that you need to know what this Expr1007 is, look at the Probe Value (green text) in the text plan and you will see that it is the result of the join. If the value passed to the INSERT at the column ID_Gender exists in the table Tab2, then that probe will return the join value; otherwise it will return NULL. So the Assert is checking the value of the search at the Tab2; if the value that is passed to the INSERT is not found  then Assert will show one exception. If the value passed to the column ID_Genders is NULL than the SQL can't show a exception, in that case it returns "0" and keeps running the query. If you run the INSERT above, the SQL will show an exception because of the "X" value, but if you change the "X" to "F" and run again, it will show an exception because of the value "4". If you change the value "4" to NULL, 1, 2 or 3 the insert will be executed without any error. Assert checking a SubQuery: The Assert operator is also used to check one subquery. As we know, one scalar subquery can't validly return more than one value: Sometimes, however, a  mistake happens, and a subquery attempts to return more than one value . Here the Assert comes into play by validating the condition that a scalar subquery returns just one value. Take the following query: INSERT INTO Tab1(ID_TipoSexo, Sexo) VALUES((SELECT ID_TipoSexo FROM Tab1), 'F')    INSERT INTO Tab1(ID_TipoSexo, Sexo) VALUES((SELECT ID_TipoSexo FROM Tab1), 'F')    |--Assert(WHERE:(CASE WHEN NOT [Pass1016] AND [Expr1015] IS NULL THEN (0) ELSE NULL END))        |--Nested Loops(Left Semi Join, PASSTHRU:([tempdb].[dbo].[Tab1].[ID_TipoSexo] IS NULL), OUTER REFERENCES:([tempdb].[dbo].[Tab1].[ID_TipoSexo]), DEFINE:([Expr1015] = [PROBE VALUE]))              |--Assert(WHERE:([Expr1017]))             |    |--Compute Scalar(DEFINE:([Expr1017]=CASE WHEN [tempdb].[dbo].[Tab1].[Sexo]<>'F' AND [tempdb].[dbo].[Tab1].[Sexo]<>'M' THEN (0) ELSE NULL END))              |         |--Clustered Index Insert(OBJECT:([tempdb].[dbo].[Tab1].[PK__Tab1__3214EC277097A3C8]), SET:([tempdb].[dbo].[Tab1].[ID_TipoSexo] = [Expr1008],[tempdb].[dbo].[Tab1].[Sexo] = [Expr1009],[tempdb].[dbo].[Tab1].[ID] = [Expr1003]))              |              |--Top(TOP EXPRESSION:((1)))              |                   |--Compute Scalar(DEFINE:([Expr1008]=[Expr1014], [Expr1009]='F'))              |                        |--Nested Loops(Left Outer Join)              |                             |--Compute Scalar(DEFINE:([Expr1003]=getidentity((1856985942),(2),NULL)))              |                             |    |--Constant Scan              |                             |--Assert(WHERE:(CASE WHEN [Expr1013]>(1) THEN (0) ELSE NULL END))              |                                  |--Stream Aggregate(DEFINE:([Expr1013]=Count(*), [Expr1014]=ANY([tempdb].[dbo].[Tab1].[ID_TipoSexo])))             |                                       |--Clustered Index Scan(OBJECT:([tempdb].[dbo].[Tab1].[PK__Tab1__3214EC277097A3C8]))              |--Clustered Index Seek(OBJECT:([tempdb].[dbo].[Tab2].[PK__Tab2__3214EC27755C58E5]), SEEK:([tempdb].[dbo].[Tab2].[ID]=[tempdb].[dbo].[Tab1].[ID_TipoSexo]) ORDERED FORWARD)  You can see from this text showplan that SQL Server as generated a Stream Aggregate to count how many rows the SubQuery will return, This value is then passed to the Assert which then does its job by checking its validity. Is very interesting to see that  the Query Optimizer is smart enough be able to avoid using assert operators when they are not necessary. For instance: INSERT INTO Tab1(ID_TipoSexo, Sexo) VALUES((SELECT ID_TipoSexo FROM Tab1 WHERE ID = 1), 'F') INSERT INTO Tab1(ID_TipoSexo, Sexo) VALUES((SELECT TOP 1 ID_TipoSexo FROM Tab1), 'F')  For both these INSERTs, the Query Optimiser is smart enough to know that only one row will ever be returned, so there is no need to use the Assert. Well, that's all folks, I see you next week with more "Operators". Cheers, Fabiano

    Read the article

facebook twitter digg google delicious technorati stumbleupon myspace wordpress linkedin gmail igoogle windows live tumblr viadeo yahoo buzz yahoo mail yahoo bookmarks favorites email print

< Previous Page | 97 98 99 100 101 102 103 104 105 106 107 108  | Next Page >