Showing posts with label happenslooking. Show all posts
Showing posts with label happenslooking. Show all posts

Wednesday, March 21, 2012

Fault tolerant distributor in merge replication and data mirroring (in SQL Server 2005/200

Cluster it. It should be a local distributor.
http://www.zetainteractive.com - Shift Happens!
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
"Josep" <jmartinez@.autec.es> wrote in message
news:OT0UAqhUIHA.5404@.TK2MSFTNGP03.phx.gbl...
> Hello!
> My customer has SQL Server 2000 with Merge Replication, implementing a
> star tipology: one publisher (also distributor) and 11 push
> subscribers.And they want to move to SQL Server 2005/2008 and have a fault
> tolerant environment.
> I've thought that probably the best is to set up data mirroring for the
> publication, but the BOL says that data mirroring is not possible for the
> distribution database. So, what's the best way to do implement a fault
> tolerant system for the distribution database? (it's not needed to be
> automatic)
> Maybe doing backups and restores quite ofter of distribution database from
> the principal to the mirror? because we have the point that in merge
> replication only changes occurs there when the replication-schema is
> changed or a new identity range is given to a subscriber (when using
> automatic identity range).
>
> Thanks in advance!
> Josep
>
>
Your English is much better than my Spanish, and IMHO extremely good!
Again I think a restore of the publisher database or the distribution
database is your best bet. If the subscribers get ahead of the
distribution database due to a restore from an earlier version of the
distribution database, the subscribers will fill in any missing data
on the publisher.
On Jan 9, 5:13 am, "Josep" <jmarti...@.autec.es> wrote:[vbcol=seagreen]
> Ok. That's the best option, but it's quite expensive and my customer doesn't
> want it. When I said fault-tolerant I didn't mean "real time recovery", I
> should better have said that the downtime recovery process should be quite
> small: less than two hours.
> As it's a Merge Replication there can be this small downtime, that's why I
> was saying doing backup/restore of distribution database... but then I don't
> know how exactly would connect the subscribers to the distribution or other
> issues that can be there and I don't know... or, if as you said, cluster is
> the only option.
> Thank you!
> Josep.
> PS: Hilary, sorry for my English :-(
> "Hilary Cotter" <hilary.cot...@.gmail.com> escribi en el mensajenews:uDoLtIiUIHA.5132@.TK2MSFTNGP02.phx.gbl. ..
>
>
>
>