December 5, 2015 at 2:43 am
Hi,
On my company I saw that people are generating snapshots with a three days interval.
We are using merge replication.
I don't understand one thing:
If we are in merge replication, the merge agent has to continuaselly track changes and update directlely the publisher database and the subscriver database every time there is a difference between them (I don't know if you can or not set the time for this agent watch).
If the snapshot agent is responsible to generate the published arthicals data and schema information, if it only generates from three to three days, how will the merge agent know that has changed from the publisher side?
Thank you
December 7, 2015 at 8:14 pm
First - once you have a subscriber setup with its initial snapshot, the snapshot agent is not required to be run.
For merge replication, each database (publisher and subscribers) have triggers on the tables being replicated that insert records into tables that are used by merge replication to sort out what records have changed. When a subscriber syncs with the publisher, the merge agent looks at these tables to determine which rows in each of the tables need to be sync'ed. It then checks the current values on those row using the rule you have defined and updates the row on each database accordingly. The rules for which row or columns are the "correct" is dependent upon the rules you define.
December 9, 2015 at 9:18 am
Ok. So, once the initial snapshot has been applyed on the subscriber it is no longer needed according to your explanation.
So, Way are they gereration new snapshots from 3 to 3 days?
What is the relevance of that?
Thank you
January 3, 2016 at 8:08 pm
There is no need to snapshots after the first one. Chances are, whoever setup that 3 day schedule did not understand the architecture of merge replication.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply