Subscribers are bound to their publishers. While merge replication does
"BATMAN" <BATMAN@discussions.microsoft.com> wrote in message
news:6FAA9D27-2BDE-4265-B58C-25BA44626DC1@microsoft.com...
> OK... A more simple way to look at it... Does SQL Server bind
> subscriptions
> to publications? After thinking about the whole repbulishing logic, the
> only
> way a subscriber could get data from another publisher/republisher is if
> the
> publication metadata, in the distribution db's, we'ere identical... And
> since republishers are independent publications created off of the source
> publication, that is not possible... So overall, to anser my own
> question, a
> subsription cannot use the topology that is diagramed out in Microsofts
> documentation, or at least the documentation is not detail enough to
> explain
> how you would implement a republisher topology against a load balanced web
> sychronization hiearchy.
>
> With that said... The only way I can see this solution working is to code
> our own logic that does round robin load balancing and eveninly distribute
> subsribers across web/sql pairs...
>
> Anyone have any thoughts?
>
> Thanks!
>
> "BATMAN" wrote:
>
>> We are using Web Synchronization against a Republisher topology...
>> Meaning
>> we are using load balanced IIS servers against republishers. Right now
>> Each
>> IIS server has a dedicated republisher pointing to a central server, not
>> across a WAN though. Our goal was mainly to create a scalable
>> environment
>> that will scale as our subscriber base grows.
>> The topology that I speak of is described in the below BOL article
>> ("Multiple IIS Systems and SQL Server Republishing" section at the bottom
>> of
>> the article.). Below is the link to that article:
>>
>>
http://msdn2.microsoft.com/en-us/library/ms151811.aspx >>
>> With that said. We use RMO to setup the pull subscription on the client
>> for
>> Web Synchronization. We setup the client inline with the subscription
>> properties necessary to use Web Synchronization. Below are the steps
>> outlined in BOL:
>>
>>
http://msdn2.microsoft.com/en-us/library/ms345207.aspx >>
>> ------------------------------------------------------------------------------------------
>> SQL Server 2005 Books Online
>> How to: Configure a Subscription to Use Web Synchronization (RMO
>> Programming)
>> ------------------------------------------------------------------------------------------
>> Set the following subscription properties:
>> 1) The ServerConnection to the Subscriber created in step 1 for
>> ConnectionContext.
>> 2) Name of the subscription database for DatabaseName.
>> 3) Name of the Publisher for PublisherName.
>> 4) Name of the publication database for PublicationDBName.
>> 5) Name of the publication for PublicationName.
>> 6) Value of true for UseWebSynchronization.
>> 7) Location of the Web server that hosts Web synchronization for
>> InternetUrl.
>> 8) Login and password for the Microsoft Windows account under which the
>> Merge Agent runs
>> 9) Value of 0 for InternetSecurityMode as well as values for
>> InternetLogin
>> and InternetPassword when using HTTP Basic Authentication to access the
>> Web
>> server using SQL Authentication credentials.
>> ------------------------------------------------------------------------------------------
>>
>> The problem is step "3) Name of the Publisher for PublisherName. " and
>> how
>> it relates to load balancing the web servers.
>>
>> So with all that said. The challenge is how do we get around having to
>> pass
>> the PublisherName which is the server/instance name when creating and
>> syncing
>> the subscriptions in a load balanced Web Synchronization/Republisher
>> topology
>> (a 1:1 Web to Republisher scenario.)?
>>
>> Thanks!
>>
>>
>>