transactional replication

[unseth]

In my system (SQL 2000) i had snapshot replication working fine.

However, I am learning about transactional replication, and I have a few questions.

1) why is there still a snapshot being made? I thought transact was supposed to just make a list of the changes that were made and then send those over?

Can I disable the snapshot and just let the transact job run? I have a feeling i'm missing something here...

2) i have a full text index on one of the tables being replicated. Snapshot was moving the entire dataset over, so the index had to be rebuilt on the subscriber machine daily. I'm hoping that using transactional replication will solve this, and the change tracking on the subscriber machine will only index the new data brought over by the transact. Is this a correct assumption?

Thanks in advance for all your help.

Jeff

 

[SQLBill]

My understanding of transactional replication is that when it is FIRST started, snapshot replication copies the database as it is right now. Then transaction replication begins and snapshot replication is not used anymore. Without the initial snapshot, transaction replication cannot take place because there's nothing for it to replicate on. (In other words, how can you change something if it doesn't exist?)

-SQLBill

 

[boing]

HI:

The first portion of Transactional replication will make a snapshot which is in essence a complete copy of the DB. Then the transactional updateing takes place either continuously or on schedule. I have found difficulty depending on the size of the database with the Snapshot agents stopping for periods of time and then restarting. Sometimes never starting back up. I have never heard an explanation for this. I have found it only on Very Large Data Bases (VLDB). Databases generally over 1 Gig. If your snapshop never finshes the transactional update will never start.

Best luck.

Boing