Geekly Articles each Day
📜 ⬆️ ⬇️

Examining DB and DBMS with T-SQL

Foreword


Greetings to you again, dear reader Habra!

When you realize your ideas, experience, as well as all the information that does not give you peace of mind in a publication, sooner or later a logical point will come to the entire previously written flow of information. This article will be different from all previously published by me with its laxity and more free style of text presentation, and also it will complete the presentation of all my accumulated experience on MS SQL Server.

This article is a supplement to the article Explore databases using T-SQL , and also briefly talks about the SRV administration database created and utility projects that are designed to help DBA MS SQL Server work.

Some useful insights for database and database research in general


To determine the size of the embedded tables, you can create the following view [inf]. [VInnerTableSize]:
')
Implementing view [inf]. [VInnerTableSize]
USE [_] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE VIEW [inf].[vInnerTableSize] as --   select object_name(p.[object_id]) as [Name] , SUM(a.[total_pages]) as TotalPages , SUM(p.[rows]) as CountRows , cast(SUM(a.[total_pages]) * 8192/1024. as decimal(18, 2)) as TotalSizeKB from sys.partitions as p inner join sys.allocation_units as a on p.[partition_id]=a.[container_id] left outer join sys.internal_tables as it on p.[object_id]=it.[object_id] where OBJECTPROPERTY(p.[object_id], N'IsUserTable')=0 group by object_name(p.[object_id]) --order by p.[rows] desc; GO

With this view, you can control the growth of system tables in order to avoid their excessive growth.

With the help of the system views [sys]. [Sql_logins] and [sys]. [Syslogins] you can get logins for cheek and screw inputs.

Also of interest are the following system views for the tasks of the Agent instance of MS SQL Server:

1) [msdb]. [Dbo]. [Sysjobactivity] - active tasks
2) [msdb]. [Dbo]. [Sysjobhistory] - task execution history
3) [msdb]. [Dbo]. [Sysjobs_view] and [msdb]. [Dbo]. [Sysjobservers] - tasks
4) [msdb]. [Dbo]. [Sysjobschedules] - task schedules
5) [msdb]. [Dbo]. [Sysjobsteps] - job steps
6) [msdb]. [Dbo]. [Sysjobstepslogs] - logging of task steps

Also, in order to know for which schedules more than one task is assigned, it is enough to create the following representation [inf]. [VScheduleMultiJobs]:

The implementation of the [inf]. [VScheduleMultiJobs] view
 USE [  ] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE view [inf].[vScheduleMultiJobs] as with sh as( SELECT schedule_id FROM [inf].[vJobSchedules] group by schedule_id having count(*)>1 ) select * from msdb.dbo.sysschedules as s where exists(select top(1) 1 from sh where sh.schedule_id=s.schedule_id) GO

This view will avoid a rash change in the schedule for one task, so as not to cause changes for another.

To obtain information about the description of database objects, you can use the extended properties (system representation [sys]. [Extended_properties] ). For convenience, you can create the following views:
1) [inf]. [VObjectDescription]:

Implementing view [inf]. [VObjectDescription]
 USE [_] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE view [inf].[vObjectDescription] as select SCHEMA_NAME(obj.[schema_id]) as SchemaName ,QUOTENAME(object_schema_name(obj.[object_id]))+'.'+quotename(obj.[name]) as ObjectName ,obj.[type] as [Type] ,obj.[type_desc] as [TypeDesc] ,ep.[value] as ObjectDescription from sys.objects as obj left outer join sys.extended_properties as ep on obj.[object_id]=ep.[major_id] and ep.[minor_id]=0 and ep.[name]='MS_Description' where obj.[is_ms_shipped]=0 and obj.[parent_object_id]=0 GO

2) Descriptions for objects that have parents — using the [inf]. [VObjectInParentDescription] view:

Implementing view [inf]. [VObjectInParentDescription]
 USE [_] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE view [inf].[vObjectInParentDescription] as select SCHEMA_NAME(obj.[schema_id]) as SchemaName ,QUOTENAME(object_schema_name(obj.[parent_object_id]))+'.'+quotename(object_name(obj.[parent_object_id])) as ParentObjectName ,QUOTENAME(object_schema_name(obj.[object_id]))+'.'+quotename(obj.[name]) as ObjectName ,obj.[type] as [Type] ,obj.[type_desc] as [TypeDesc] ,ep.[value] as ObjectDescription from sys.all_objects as obj left outer join sys.extended_properties as ep on obj.[parent_object_id]=ep.[major_id] and ep.[minor_id]=obj.[object_id] and ep.[name]='MS_Description' where obj.[is_ms_shipped]=0 and obj.[parent_object_id]<>0 GO

3) Parameter descriptions - using the [inf]. [VParameterDescription] view:

Implementing view [inf]. [VParameterDescription]
 USE [_] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE view [inf].[vParameterDescription] as select SCHEMA_NAME(obj.[schema_id]) as SchemaName ,QUOTENAME(object_schema_name(obj.[object_id]))+'.'+quotename(object_name(obj.[object_id])) as ParentObjectName ,p.[name] as ParameterName ,obj.[type] as [Type] ,obj.[type_desc] as [TypeDesc] ,ep.[value] as ParameterDescription from sys.parameters as p inner join sys.objects as obj on p.[object_id]=obj.[object_id] left outer join sys.extended_properties as ep on obj.[object_id]=ep.[major_id] and ep.[minor_id]=p.[parameter_id] and ep.[name]='MS_Description' where obj.[is_ms_shipped]=0 GO

4) Table column descriptions using the [inf]. [VColumnTableDescription] view:

Implementing view [inf]. [VColumnTableDescription]
 USE [_] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE VIEW [inf].[vColumnTableDescription] as select SCHEMA_NAME(t.schema_id) as SchemaName ,QUOTENAME(object_schema_name(t.[object_id]))+'.'+quotename(t.[name]) as TableName ,c.[name] as ColumnName ,ep.[value] as ColumnDescription from sys.tables as t inner join sys.columns as c on c.[object_id]=t.[object_id] left outer join sys.extended_properties as ep on t.[object_id]=ep.[major_id] and ep.[minor_id]=c.[column_id] and ep.[name]='MS_Description' where t.[is_ms_shipped]=0; GO

5) View column descriptions using the [inf]. [VColumnViewDescription] view:

Implementing view [inf]. [VColumnViewDescription]
 USE [_] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE VIEW [inf].[vColumnViewDescription] as select SCHEMA_NAME(t.schema_id) as SchemaName ,QUOTENAME(object_schema_name(t.[object_id]))+'.'+quotename(t.[name]) as TableName ,c.[name] as ColumnName ,ep.[value] as ColumnDescription from sys.views as t inner join sys.columns as c on c.[object_id]=t.[object_id] left outer join sys.extended_properties as ep on t.[object_id]=ep.[major_id] and ep.[minor_id]=c.[column_id] and ep.[name]='MS_Description' where t.[is_ms_shipped]=0; GO

6) DB schema descriptions using the [inf]. [VSchemaDescription] view:

Implementing the view [inf]. [VSchemaDescription]
 USE [_] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE VIEW [inf].[vSchemaDescription] as select SCHEMA_NAME(t.schema_id) as SchemaName --,QUOTENAME(object_schema_name(t.[object_id]))+'.'+quotename(t.[name]) as TableName ,ep.[value] as SchemaDescription from sys.schemas as t left outer join sys.extended_properties as ep on t.[schema_id]=ep.[major_id] and ep.[minor_id]=0 and ep.[name]='MS_Description' GO

To add or edit extended properties for documenting database objects, it is better to use third-party utilities (for example, I use dbForge ).
However, it can also be done with the following queries:

Examples of creating descriptions for database objects
 --     @ObjectID - dbo.GetPlansObject --       EXECUTE sp_addextendedproperty @name = N'MS_Description', @value = N' ', @level0type = N'SCHEMA', @level0name = N'dbo', @level1type = N'FUNCTION', @level1name = N'GetPlansObject', @level2type = N'PARAMETER', @level2name = N'@ObjectID'; --    - dbo.GetPlansObject --     ,  EXEC sys.sp_addextendedproperty @name=N'MS_Description', @value=N'    ', @level0type=N'SCHEMA', @level0name=N'dbo', @level1type=N'FUNCTION', @level1name=N'GetPlansObject'; --     inf.vColumnTableDescription --     EXEC sys.sp_addextendedproperty @name=N'MS_Description', @value=N'  ', @level0type=N'SCHEMA', @level0name=N'inf', @level1type=N'VIEW', @level1name=N'vColumnTableDescription'; --     TEST_GUID  dbo.TABLE --      EXEC sys.sp_addextendedproperty @name=N'MS_Description', @value=N'  ()', @level0type=N'SCHEMA', @level0name=N'dbo', @level1type=N'TABLE', @level1name=N'TEST', @level2type=N'COLUMN', @level2name=N'TEST_GUID'; --     rep EXEC sys.sp_addextendedproperty @name=N'MS_Description', @value=N'  rep    ' , @level0type=N'SCHEMA', @level0name=N'rep'; --     EXEC sys.sp_addextendedproperty @name=N'MS_Description', @value=N'      MS SQL Server 2016-2017 (     MS SQL Server 2012-2014).      MS SQL Server 2012           .    .   inf.InfoAgentJobs.';

In order to change or delete the description, it suffices to use the stored procedures sp_updateextendedproperty and sp_dropextendedproperty, respectively.

The following system views will also be useful in the study of the entire DBMS:

1) [sys]. [Dm_os_performance_counters] - values ​​of performance counters

2) [sys]. [Dm_os_schedulers] - Task Schedulers

3) [sys]. [Configurations] - configuration information

4) to match the session identifiers with the Windows thread identifiers, you can create the following representation [inf]. [VSessionThreadOS]:

The implementation of the [inf]. [VSessionThreadOS] view
 USE [  ] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO create view [inf].[vSessionThreadOS] as /*   ,       Windows.         Windows.     ,        . */ SELECT STasks.session_id, SThreads.os_thread_id FROM sys.dm_os_tasks AS STasks INNER JOIN sys.dm_os_threads AS SThreads ON STasks.worker_address = SThreads.worker_address WHERE STasks.session_id IS NOT NULL; GO

5) to learn about problems with the number of tempdb database files, you can create the following representation [inf]. [VServerProblemInCountFilesTempDB]:

Implementing view [inf]. [VServerProblemInCountFilesTempDB]
 USE [  ] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO create view [inf].[vServerProblemInCountFilesTempDB] as /* http://sqlcom.ru/dba-tools/tempdb-in-sql-server-2016/          tempdb.     latch    PFS, GAM, SGAM    tempdb.           «Is Not PFS, GAM, or SGAM page»,          tempdb */ Select session_id, wait_type, wait_duration_ms, blocking_session_id, resource_description, ResourceType = Case When Cast(Right(resource_description, Len(resource_description) - Charindex(':', resource_description, 3)) As Int) - 1 % 8088 = 0 Then 'Is PFS Page' When Cast(Right(resource_description, Len(resource_description) - Charindex(':', resource_description, 3)) As Int) - 2 % 511232 = 0 Then 'Is GAM Page' When Cast(Right(resource_description, Len(resource_description) - Charindex(':', resource_description, 3)) As Int) - 3 % 511232 = 0 Then 'Is SGAM Page' Else 'Is Not PFS, GAM, or SGAM page' End From sys.dm_os_waiting_tasks Where wait_type Like 'PAGE%LATCH_%' And resource_description Like '2:%' GO

6) to learn about problems with the data writing time in the tempdb database, you can create the following representation [srv]. [VStatisticsIOInTempDB]:

Implementing view [srv]. [VStatisticsIOInTempDB]
 USE [  ] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO create view [srv].[vStatisticsIOInTempDB] as /*     (avg_write_stall_ms)  5 ,      .  5  10  —  .  10  —  ,    ,    -     https://minyurov.com/2016/07/24/mssql-tempdb-opt/ */ SELECT files.physical_name, files.name, stats.num_of_writes, (1.0 * stats.io_stall_write_ms / stats.num_of_writes) AS avg_write_stall_ms, stats.num_of_reads, (1.0 * stats.io_stall_read_ms / stats.num_of_reads) AS avg_read_stall_ms FROM sys.dm_io_virtual_file_stats(2, NULL) as stats INNER JOIN master.sys.master_files AS files ON stats.database_id = files.database_id AND stats.file_id = files.file_id WHERE files.type_desc = 'ROWS' GO

7) for the convenience of displaying information about the latest backup copies of all databases, you can create the following representation [inf]. [VServerLastBackupDB]:

Implementing view [inf]. [VServerLastBackupDB]
 USE [  ] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE VIEW [inf].[vServerLastBackupDB] as with backup_cte as ( select bs.[database_name], backup_type = case bs.[type] when 'D' then 'database' when 'L' then 'log' when 'I' then 'differential' else 'other' end, bs.[first_lsn], bs.[last_lsn], bs.[backup_start_date], bs.[backup_finish_date], cast(bs.[backup_size] as decimal(18,3))/1024/1024 as BackupSizeMb, rownum = row_number() over ( partition by bs.[database_name], type order by bs.[backup_finish_date] desc ), LogicalDeviceName = bmf.[logical_device_name], PhysicalDeviceName = bmf.[physical_device_name], bs.[server_name], bs.[user_name] FROM msdb.dbo.backupset bs INNER JOIN msdb.dbo.backupmediafamily bmf ON [bs].[media_set_id] = [bmf].[media_set_id] ) select [server_name] as [ServerName], [database_name] as [DBName], [user_name] as [USerName], [backup_type] as [BackupType], [backup_start_date] as [BackupStartDate], [backup_finish_date] as [BackupFinishDate], [BackupSizeMb], --   [LogicalDeviceName], [PhysicalDeviceName], [first_lsn] as [FirstLSN], [last_lsn] as [LastLSN] from backup_cte where rownum = 1; GO

8) a similar representation [inf]. [VServerBackupDB] can be created to obtain information about all backup copies:
Implementing view [inf]. [VServerBackupDB]
 USE [  ] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE VIEW [inf].[vServerBackupDB] as with backup_cte as ( select bs.[database_name], backup_type = case bs.[type] when 'D' then 'database' when 'L' then 'log' when 'I' then 'differential' else 'other' end, bs.[first_lsn], bs.[last_lsn], bs.[backup_start_date], bs.[backup_finish_date], cast(bs.[backup_size] as decimal(18,3))/1024/1024 as BackupSizeMb, LogicalDeviceName = bmf.[logical_device_name], PhysicalDeviceName = bmf.[physical_device_name], bs.[server_name], bs.[user_name] FROM msdb.dbo.backupset bs INNER JOIN msdb.dbo.backupmediafamily bmf ON [bs].[media_set_id] = [bmf].[media_set_id] ) select [server_name] as [ServerName], [database_name] as [DBName], [user_name] as [USerName], [backup_type] as [BackupType], [backup_start_date] as [BackupStartDate], [backup_finish_date] as [BackupFinishDate], [BackupSizeMb], --   [LogicalDeviceName], [PhysicalDeviceName], [first_lsn] as [FirstLSN], [last_lsn] as [LastLSN] from backup_cte; GO

9) you can also improve the performance on expectations statistics (from the article SQL Server expectations statistics or please tell me where it hurts ) to remove the output duplicate lines in the form of [inf]. [VWaits]:

Implementing the view [inf]. [VWaits]
 USE [  ] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE view [inf].[vWaits] as WITH [Waits] AS (SELECT [wait_type], --   [wait_time_ms] / 1000.0 AS [WaitS],--      .    signal_wait_time_ms ([wait_time_ms] - [signal_wait_time_ms]) / 1000.0 AS [ResourceS],--        signal_wait_time_ms [signal_wait_time_ms] / 1000.0 AS [SignalS],--           [waiting_tasks_count] AS [WaitCount],--   .         100.0 * [wait_time_ms] / SUM ([wait_time_ms]) OVER() AS [Percentage], ROW_NUMBER() OVER(ORDER BY [wait_time_ms] DESC) AS [RowNum] FROM sys.dm_os_wait_stats WHERE [wait_type] NOT IN ( N'BROKER_EVENTHANDLER', N'BROKER_RECEIVE_WAITFOR', N'BROKER_TASK_STOP', N'BROKER_TO_FLUSH', N'BROKER_TRANSMITTER', N'CHECKPOINT_QUEUE', N'CHKPT', N'CLR_AUTO_EVENT', N'CLR_MANUAL_EVENT', N'CLR_SEMAPHORE', N'DBMIRROR_DBM_EVENT', N'DBMIRROR_EVENTS_QUEUE', N'DBMIRROR_WORKER_QUEUE', N'DBMIRRORING_CMD', N'DIRTY_PAGE_POLL', N'DISPATCHER_QUEUE_SEMAPHORE', N'EXECSYNC', N'FSAGENT', N'FT_IFTS_SCHEDULER_IDLE_WAIT', N'FT_IFTSHC_MUTEX', N'HADR_CLUSAPI_CALL', N'HADR_FILESTREAM_IOMGR_IOCOMPLETION', N'HADR_LOGCAPTURE_WAIT', N'HADR_NOTIFICATION_DEQUEUE', N'HADR_TIMER_TASK', N'HADR_WORK_QUEUE', N'KSOURCE_WAKEUP', N'LAZYWRITER_SLEEP', N'LOGMGR_QUEUE', N'ONDEMAND_TASK_QUEUE', N'PWAIT_ALL_COMPONENTS_INITIALIZED', N'QDS_PERSIST_TASK_MAIN_LOOP_SLEEP', N'QDS_CLEANUP_STALE_QUERIES_TASK_MAIN_LOOP_SLEEP', N'REQUEST_FOR_DEADLOCK_SEARCH', N'RESOURCE_QUEUE', N'SERVER_IDLE_CHECK', N'SLEEP_BPOOL_FLUSH', N'SLEEP_DBSTARTUP', N'SLEEP_DCOMSTARTUP', N'SLEEP_MASTERDBREADY', N'SLEEP_MASTERMDREADY', N'SLEEP_MASTERUPGRADED', N'SLEEP_MSDBSTARTUP', N'SLEEP_SYSTEMTASK', N'SLEEP_TASK', N'SLEEP_TEMPDBSTARTUP', N'SNI_HTTP_ACCEPT', N'SP_SERVER_DIAGNOSTICS_SLEEP', N'SQLTRACE_BUFFER_FLUSH', N'SQLTRACE_INCREMENTAL_FLUSH_SLEEP', N'SQLTRACE_WAIT_ENTRIES', N'WAIT_FOR_RESULTS', N'WAITFOR', N'WAITFOR_TASKSHUTDOWN', N'WAIT_XTP_HOST_WAIT', N'WAIT_XTP_OFFLINE_CKPT_NEW_LOG', N'WAIT_XTP_CKPT_CLOSE', N'XE_DISPATCHER_JOIN', N'XE_DISPATCHER_WAIT', N'XE_TIMER_EVENT') ) , ress as ( SELECT [W1].[wait_type] AS [WaitType], CAST ([W1].[WaitS] AS DECIMAL (16, 2)) AS [Wait_S],--      .    signal_wait_time_ms CAST ([W1].[ResourceS] AS DECIMAL (16, 2)) AS [Resource_S],--        signal_wait_time_ms CAST ([W1].[SignalS] AS DECIMAL (16, 2)) AS [Signal_S],--           [W1].[WaitCount] AS [WaitCount],--   .         CAST ([W1].[Percentage] AS DECIMAL (5, 2)) AS [Percentage], CAST (([W1].[WaitS] / [W1].[WaitCount]) AS DECIMAL (16, 4)) AS [AvgWait_S], CAST (([W1].[ResourceS] / [W1].[WaitCount]) AS DECIMAL (16, 4)) AS [AvgRes_S], CAST (([W1].[SignalS] / [W1].[WaitCount]) AS DECIMAL (16, 4)) AS [AvgSig_S] FROM [Waits] AS [W1] INNER JOIN [Waits] AS [W2] ON [W2].[RowNum] <= [W1].[RowNum] GROUP BY [W1].[RowNum], [W1].[wait_type], [W1].[WaitS], [W1].[ResourceS], [W1].[SignalS], [W1].[WaitCount], [W1].[Percentage] HAVING SUM ([W2].[Percentage]) - [W1].[Percentage] < 95 -- percentage threshold ) SELECT [WaitType] ,MAX([Wait_S]) as [Wait_S] ,MAX([Resource_S]) as [Resource_S] ,MAX([Signal_S]) as [Signal_S] ,MAX([WaitCount]) as [WaitCount] ,MAX([Percentage]) as [Percentage] ,MAX([AvgWait_S]) as [AvgWait_S] ,MAX([AvgRes_S]) as [AvgRes_S] ,MAX([AvgSig_S]) as [AvgSig_S] FROM ress group by [WaitType]; GO

SRV DB for database administration and DBMS in general


Until the moment of writing my articles, I created the SRV DB, which was modified and supplemented with the experience and knowledge gained.
Other utility projects were also developed to help the database administrator on C # .NET.

Access to projects here .
At the root is the “Description” file, where each project is briefly described.
These solutions are open and distributed freely.

Also thanks to you, dear Habr readers, who left feedback in the form of comments and messages, we managed to improve the project on the SRV database. Thank you very much for that!

But it is important to note that existing approaches and solutions described in external sources should be carefully analyzed, since These methods may not be suitable for your task. It is necessary to pay special attention to the difference in the parameters and conditions of your task from the task, which is solved in the example: load, amount of information processed, frequency, specificity of a business problem, etc. For example, a procedure that works for 40 minutes is suitable for a call once a day, but if the process needs to be started with a higher frequency, then this solution may not be suitable.

Having found your unique approach to a specific task, do not forget to share it! Thus, you will add to the “global knowledge base”, which facilitates the search for solutions and ideas for new tasks.

Results


Some more useful system views of MS SQL Server were considered, including for self-documentation in the form of extended properties.

Reflections and ideas


As you have already noticed, MS SQL Server already sufficiently supports NoSQL in the form of graph tables (with MS SQL Server 2017 ) and document-oriented data (XML, and with MS SQL Server 2016 and JSON ).

However, as Edgar Frank Codd (by source [1]) noted in the 70s of the 20th century, not a simple relationship can be considered in the relational model. That is, you can both embed one table into another, and inherit from one table to another (I remind you that a table is a relation in a relational model). Table inheritance is implemented in some DBMS, for example, in the same PostgreSQL . But I have not seen the implementation of investments. If we implement both attachments and inheritance, as well as lay the mechanism for processing these complex relationships, we get a DBMS that summarizes the JSON and XML formats and fully covers the so-called NoSQL technology (analogue of operator overload in programming languages, but in DBMS indices, aggregation, statistics , service and t d for such a relationship). Moreover, it will probably cover all other data models to a sufficient degree, although it will be handled by the declarative language of SQL queries with some specific extensions and definitions for complex relationships.

Seeing how quickly MS SQL Server is developing, it is worth hoping that someday it will come to the realization of complex relationships and cover all their varieties. And then the wishes and foresight of the creator of relational algebra will be brought to life, and in the relational model, specialists will discover completely different aspects of creating various information databases and data warehouses.

Sources:


"" High load applications. Programming, scaling, support ”, SPb .: Peter, 2018 Kleppman M. [1]
" SQL Server wait statistics or please tell me where it hurts
» Investigate databases with T-SQL
» SQL documentation
» Tempdb Improvements in SQL Server 2016
» Temporary DB Optimization (tempdb)
» T-SQL formatting standard
» Utilities for MS SQL Server DBA
» DbForge
PostgreSQL (inheritance)
» MS SQL Server 2017 (columns)
» JSON in MS SQL Server 2016-2017

Source: https://habr.com/ru/post/350388/

More articles:


All Articles