By John Jeheimer


2013-06-05 16:55:48 8 Comments

I have 8 individual SQL Server 2008 R2 machines, each hosting 1 database. Each database has an identical table structure and schema, and entirely unique data.

I would like to establish a reporting server (may be 2008 or 2012), that consolidates the rows from selected tables across the 8 source servers into a single instance of those tables on the reporting server. This is one-way replication (no changes will be made to the reporting server). I will need to replicate changes from the source databases with relatively low latency (say 20-30 seconds).

Furthermore, I'd like to find methods to achieve this with as little impact to the source servers as is practical. 3rd-party agents, triggers, or schema mods to those servers are difficult in my environment.

My questions:

  • What are promising architectures and technologies for achieving this goal?
  • I've looked at SQL Server Merge Replication, but I am concerned about latency. Is this an appropriate technology for this goal?
  • Are there many-to-one architectures for transactional replication?
  • Should I be looking at 1-to-1 replication into 8 databases on my reporting server, followed by some custom merge function (a 2-step replication)?

Thanks, John

1 comments

@Kin Shah 2013-06-06 20:58:04

I have 8 individual SQL Server 2008 R2 machines, each hosting 1 database. Each database has an identical table structure and schema, and entirely unique data. I would like to establish a reporting server (may be 2008 or 2012), that consolidates the rows from selected tables across the 8 source servers into a single instance of those tables on the reporting server. This is one-way replication (no changes will be made to the reporting server). I will need to replicate changes from the source databases with relatively low latency (say 20-30 seconds).

You can achieve this with Transactional replication. Below is how you can do it.

Note : You have to change slightly your table schema to achieve this as you have to uniquely identify that rows when you are replicating to the subscriber. As a prerequisite of T-Rep you need to have tables with PK defined.

Below is your sample table on Publisher servers that is on all your 8 servers that you want to consolidate rows on the reporting server :

CREATE TABLE Products
(
ProductID INT not null,
ProductName VARCHAR(25),
ServerName sysname default @@servername not null -- this is to identify which row is from which server ; probably add this using Alter column
)
GO
ALTER TABLE Products
ADD CONSTRAINT pk_Product_ID_ServerName PRIMARY KEY (ProductID)

On the subscriber server, you need to create the same table but with different PK to uniquely identify the rows at subscriber (not doing so, T-Rep is going to fail with PK violation - I am assuming that you cannot modify the PK structure on live PRODUCTION rather its better to modify at the subscriber)

CREATE TABLE Products
(
ProductID INT not null,
ProductName VARCHAR(25),
ServerName sysname default @@servername not null
);
GO

ALTER TABLE Products
ADD CONSTRAINT pk_Product_ID_ServerName PRIMARY KEY (ProductID,ServerName)

Below script will help you setting up T-Rep, just change the databasename, destination server name along with object name.

-- Enabling the replication database
use master
exec sp_replicationdboption @dbname = N'repl1', @optname = N'publish', @value = N'true'
GO

exec [repl1].sys.sp_addlogreader_agent @job_login = null, @job_password = null, @publisher_security_mode = 1
GO
exec [repl1].sys.sp_addqreader_agent @job_login = null, @job_password = null, @frompublisher = 1
GO
-- Adding the transactional publication
use [repl1]
exec sp_addpublication @publication = N'repl1_2005', @description = N'Transactional publication of database ''repl1'' from Publisher ''server_name\SQL2005''.', @sync_method = N'concurrent', @retention = 0, @allow_push = N'true', @allow_pull = N'true', @allow_anonymous = N'false', @enabled_for_internet = N'false', @snapshot_in_defaultfolder = N'true', @compress_snapshot = N'false', @ftp_port = 21, @ftp_login = N'anonymous', @allow_subscription_copy = N'false', @add_to_active_directory = N'false', @repl_freq = N'continuous', @status = N'active', @independent_agent = N'true', @immediate_sync = N'false', @allow_sync_tran = N'false', @autogen_sync_procs = N'false', @allow_queued_tran = N'false', @allow_dts = N'false', @replicate_ddl = 1, @allow_initialize_from_backup = N'false', @enabled_for_p2p = N'false', @enabled_for_het_sub = N'false'
GO


exec sp_addpublication_snapshot @publication = N'repl1_2005', @frequency_type = 1, @frequency_interval = 0, @frequency_relative_interval = 0, @frequency_recurrence_factor = 0, @frequency_subday = 0, @frequency_subday_interval = 0, @active_start_time_of_day = 0, @active_end_time_of_day = 235959, @active_start_date = 0, @active_end_date = 0, @job_login = null, @job_password = null, @publisher_security_mode = 1
exec sp_grant_publication_access @publication = N'repl1_2005', @login = N'sa'
GO
exec sp_grant_publication_access @publication = N'repl1_2005', @login = N'NT AUTHORITY\SYSTEM'
GO
exec sp_grant_publication_access @publication = N'repl1_2005', @login = N'BUILTIN\Administrators'
GO
exec sp_grant_publication_access @publication = N'repl1_2005', @login = N'server_name\SQLServer2005SQLAgentUser$server_name$SQL2005'
GO
exec sp_grant_publication_access @publication = N'repl1_2005', @login = N'server_name\SQLServer2005MSSQLUser$server_name$SQL2005'
GO
exec sp_grant_publication_access @publication = N'repl1_2005', @login = N'distributor_admin'
GO

-- Adding the transactional articles
use [repl1]
exec sp_addarticle @publication = N'repl1_2005', @article = N'Products', @source_owner = N'dbo', @source_object = N'Products', @type = N'logbased', @description = N'', @creation_script = N'', @pre_creation_cmd = N'none', @schema_option = 0x000000000803509F, @identityrangemanagementoption = N'none', @destination_table = N'Products', @destination_owner = N'dbo', @status = 24, @vertical_partition = N'false', @ins_cmd = N'CALL [sp_MSins_dboProducts]', @del_cmd = N'CALL [sp_MSdel_dboProducts]', @upd_cmd = N'SCALL [sp_MSupd_dboProducts]'
GO

-- Adding the transactional subscriptions
use [repl1]
exec sp_addsubscription @publication = N'repl1_2005', @subscriber = N'server_name\SQL2008R2', @destination_db = N'repl123', @subscription_type = N'Push', @sync_type = N'automatic', @article = N'all', @update_mode = N'read only', @subscriber_type = 0
exec sp_addpushsubscription_agent @publication = N'repl1_2005', @subscriber = N'server_name\SQL2008R2', @subscriber_db = N'repl123', @job_login = null, @job_password = null, @subscriber_security_mode = 1, @frequency_type = 64, @frequency_interval = 1, @frequency_relative_interval = 1, @frequency_recurrence_factor = 0, @frequency_subday = 4, @frequency_subday_interval = 5, @active_start_time_of_day = 0, @active_end_time_of_day = 235959, @active_start_date = 0, @active_end_date = 0, @dts_package_location = N'Distributor'
GO

Couple of points to note :

In sp_addsubscription make sure that @sync_type = N'automatic'

And the article properties should be set to :

enter image description here

So finally, you can have rows consolidated from all (in my case 3 servers) as below :

enter image description here

So in summary,

  • Use T-Rep.
  • Add an additional column to the existing Publisher databases e.g. serverName to uniquely identify the rows at the subscriber.
  • Create table on Subscriber having PK included as ServerName.

  • Create replication of the tables with @sync_type = N'automatic' and Article property set to "Keep existing object unchanged".

  • Run snapshot agent.

  • Check the consolidated data on the subscriber.

@Kin Shah 2013-06-07 21:24:45

@JohnJeheimer Glad that it helped. If you feel that the answer answered your question, please upvote/mark as answer.

Related Questions

Sponsored Content

3 Answered Questions

[SOLVED] Best solution for PostgreSQL partial table replication

1 Answered Questions

Merge Replication: Updates to Schema

1 Answered Questions

2 Answered Questions

[SOLVED] Oneway sync of two SQL Servers via FTP

2 Answered Questions

[SOLVED] Massive peer-to-peer replication topology?

Sponsored Content