July 31, 2007 at 8:35 pm
Thanks! in advance for the Suggestions or ideas!
I need to replicate few tables for every 10 minutes from a production database and would like to implment this by setting up Transactional replication only, but unfortunately not all the tables have primary keys.
Can any one suggest workaround for this scenario?
Workaround: what I thought is... we can create an "id" column in each table that doesn't have a primary key and then we can use these tables for transaction replication. Is this do-able??? Does this work???
Regards.
August 1, 2007 at 2:20 am
Why not setup asynchronous mirroring? There will no no impact on your prod box because it's async, and your destination server will always be up to date.
Adding a PK to your tables would allow replication but how would that affect the rest of the app? inserts, selects etc? You might end up breaking things.
You could also use SSIS to do it. Would be really simple to schedule every 10 minutes.
Cheers,CrispinI can't die, there are too many people who still have to meet me!It's not a bug, SQL just misunderstood me!
August 1, 2007 at 2:30 am
I struggle with the concept of a relational database with primary keys - an identity column does not replicate either, as such. remember a primary key does not have to be clustered so a unique key ( without nulls ) could be designated PK.
what you want to do with the "replicated" table will decide your route - mirroring doesn't allow access to the table unless you then apply a snapshot - might not be what you want. Don't mix HA and data availablity.
[font="Comic Sans MS"]The GrumpyOldDBA[/font]
www.grumpyolddba.co.uk
http://sqlblogcasts.com/blogs/grumpyolddba/
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply