Good morning,
We currently have a large database which we are partitioning,by year, into multiple filegroups.
From this, we replicate a subset of the main data table (omitting any varbinary(max) columns) to a separate server for reporting purposes.
The replication only started in the last 9 months, so this is a new situation for us.
What we need to do is add a new filegroup/file to the main database and also alter the partition scheme and partition function. That in itself is easy enough and we have no problem with that. I noticed that when I altered the partition function, a lot of data was moved from one filegroup to another. I'm not entierly sure that this is by the original DBA's design, or if the developers didn't quite understand what they were doing when designing the app itself.
My concern is what will happen at the subscriber.
I am worried that the addition of a new FG will either break replication, or that this will work fine, but all the t-log of data being moved from FG1 to FG2 will need to be replayed on the reporting server. (this is about 1 million rows).
Because we're only moving a subset of the main table, I am hoping that, should the t-log be replayed, that this wont take too long (it only takes an age on the main database because we're moving imagefiles).
Does anybody have experience with such a situation and are they able to give me a little insight as to what I can expect to happen?
Regards