In merge we want to check that the data is in sync. This is done by setting up a validation:
Checking the rowcounts and the checksums is the only surefire way to get a comparison:
Nothing happens until we run the merge agent and synchronize. At that point if the data is out of sync the job fails and we’ll get an error in the merge agent history.
The best thing to do is to have a separate job which regularly triggers the validation. I have one which runs each Monday morning and has the following type of text in it:
exec sp_validatemergepublication @publication= ‘pubTestMerge’, @level = 3
As a final point if the merge agent is run again afterwards, there is a successful sync and no mention of the invalid data.
So – perhaps the support team re-run the agent and the important info is lost. This sort of thing is done quite often because the merge agent might fail because of an innocuous network issue. Just something to be aware of.