Basic replication question--which server does the work?

[jsteph]

Hi all,
If I have Server A as a publisher and Server B as as subscriber, is Server A executing read of the log and performaing an update to Server B, or is Server B reading the log of Server A and updating itself? Or is it a configurable choice?

I'm just trying to get an idea of who's cpu cycles are being used for most of this work. I'm assuming disk i/o is moot because regardless of what machine is doing the thinking, the A log has to be read and the B table has to get written.

Thanks,
Jim

 

[dhulbert]

Server A is the Publisher and Server B is the Subscriber but there is also a Distributor which I believe is where the real leg work is done, it all depends on where you configure the distibution database.
I think the usual recommendation is to not use the publisher as a distributor.

I love deadlines. I like the whooshing sound they make as they fly by
Douglas Adams
(1952-2001)

 

[jsteph]

Thanks...Yes, I looked deeper and set it up in my test system and it gives the option of "push" or "pull", so it does give me a chance to set that.
Thanks,
--Jim