October 4, 2005 at 10:17 am
I set up an anonymous transactional pull subscription using ftp on two servers that are not in production yet. It seems to work OK, but I have some basic questions. Here is my setup:
Server A: Publisher and Distributor
Recreates snapshot daily at 4 am.
Server B: Subscriber: Anonomous ftp pull subscription with immediate updates.
Q. What is the flow of changes from Server A to Server B?
Publisher on Server A to Distributor on Server A to Server B?
OR
Publisher on Server A to Server B.
Q. What are the pros and cons of increasing the frequency of the snapshot publication?
October 5, 2005 at 3:02 pm
If I understand you correctly, than these comments come to mind:
" ... Recreates snapshot daily at 4 am. ..."
If you're doing transactional replication, you only need to run the snapshot once, not every day. After the snapshot is applied to the subscriber, you only bring over changes to tables, not the entire table. You would reinitialize the subscriber & re-do the snapshot if something happened at the subscriber to get data out of sync.
" .... Anonomous ftp pull subscription with immediate updates. ... "
"immediate updating" as defined in the publication, refers to allowing updates at the Subscriber, Server B, to send those updates back to the Publisher, Server A. This is used if a table can be updated at both the Publlisher & Subscriber & you want to keep them both in sync. If you're just sending data from A to B, then you don't need "immediate updates".
"... Q. What are the pros and cons of increasing the frequency of the snapshot publication? ..."
I take it that you're doing snapshot replication here (not transactional), separate from your other question ? You run snapshot replication as frequently as you desire the subscriber to be refreshed. However frequently you run the snaphot agent at the publisher is how frequently you need to run the agent at the subscriber to bring it over. The "pro" for frequent snapshot is more current data. The "con" is server resources, and the amount of time that your tables are locked. We do once a day snapshots for smaller tables that don't change often (product codes, state codes...), and trans replication every 15 minutes for those large tables that users want to get current data from.
I hope this helps ... Anyone ... feel free to jump in if I've misunderstood or misstated anything.
October 6, 2005 at 6:46 am
Thanks. That's the info I was looking for. I'm sorry I wasn't more clear in my description. I'm using transactional pull replication from a production database server to a read-only database server . I'm not pushing any updates back to the publisher. I wanted to verify what I seemed to being seeing: Transactional replication after the first sync was not making use of the snapshot. So there's no need to regenerate it again. Unless, as you say, the data gets out of sync for some reason.
Thanks again for your help.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply