Dear Brain,
If i understand ur questation then i suggest that doing replication is not
the best best thing and it take lot of time for replication and as u
mentioned u r going for 2 tier and if replication fails then it is a problem
for u.
So i tell u i had faced the same problem and i created a job and a cursor in
which i update all my tables every 2 hour mintues in the cursor i mentioned
that when the updation is complete in one database then it shift to other
changing the ip address of the server.
U can also update all the datebase server at a time. I had 1500 tables and 8
remotes location and it take only one hour to update.
if i had did wrong then feel free to tell me to update my knowledge.
from sufian
[quoted text, click to view] "Brian Reuter" wrote:
> I'm looking for a little help in preparing a disaster recovery staragey for
> merge replication. We are desigining a replication topology that includes
> publishers, named Republishers and several hundred anonymous merge
> subscribers.
>
> The "top tier" publisher will only talk to named subscribers, so its
> recovery via a restore and then synchronizing with the named subscribers
> should be fine. The problem I see is with the "tier 2" (republisher with
> anonymous subscriptions to it) may not be recoverable according to the
> excerpt below from the books online.
>
> The situation would be Anonymous sub updates a record, it is replicated to
> the "tier 2" box. The "Tier 2" box dies and is restored from several days
> earlier. This "tier 2" republisher is synchonized with the "top tier"
> publisher bringing it fairly current. The Anonymous subscriber now
> synchronizes again with the restore tier 2 box.
>
> Does it resend it's update? What specifically is missing on anoymous
> subscribers that will cause an issue?
>
> Any insight is appreciated!!!
>
>
> "Caution If you synchronize with a subscription database, you must
> synchronize with one that has a global subscription (that is, a subscription
> having an assigned priority value) to guarantee correct convergence behavior.
> Do not synchronize the publication database with a subscription database that
> has an anonymous subscription. Because anonymous subscriptions do not have
> enough meta data to apply changes to the publication database, such
> synchronization could lead to the non-convergence of data."
>
