February 15, 2008 at 1:27 pm
Hi All
We have snapshot replication implemented in our environment with a snapshot agent running once a day. We are planning to implement transactional replication instead of snapshot. While planning for the change there are few questions we came across.
1. Is there an option for implementing Transactional replication for tables without primary keys ( By
default it can't be, but we need to replicate few tables without primary keys.)
2. Does snapshot replication truncate the tables at subscriber databases before applying snapshot?
3. My understanding is that the distribution agent in snapshot replication moves snapshot after the snapshot agent runs. If it the case, why the action time of each distribution agent is shown as current server time ( in EM against each agent ), though the snapshot agent is running once in a day? Please correct me if I am wrong.
Thanks
RM
February 28, 2008 at 1:34 am
Hi
1. The primary key is used by the transaction process in replication to track records. Are the tables with no primary keys being updated regularly or are they rebuilt everyday. If they are updated with updates/inserts/deletes how do you track uniqueness. If the tables are rebuilt every day and you have them in transactional replication, this will generate a lot of replication traffic i.e. you can only use delete not truncate to clear the tables and all these delete commands will be replicated.
2. If I remember correctly you can either truncate the tables or apply a completely new table/indexes and data
3. The Distribution agent will show the last time it was run. The Agent can service more than one publication so if you have transactional replication running through the day the time will be updated. You are correct the agent will not ditribute the snapshot of data unitl the snapshot agent is run and creates a new set.
HTH
GRaeme
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply