Hello ,
I am reading all the texts again and again, still not able to understand
what to really do.
Database we replicate to notebooks potentially contains a lot of data.
I may understand what to do, if I want to alter column using method you call
'altering the table in-place.' , and usage of sp_repladdcolumn.
Just what to do if I need to alter application trigger for this table or
add/drop index? Do I understand well, that I have to use 'altering
subscription process' with sp_dropmergesubscription, sp_dropmergearticle,
change...., .....
Second question:
Paul wrote on
www.replicationanswers.com , 'If we drop the subscription
entirely including all other articles (sp_dropmergesubscription), then try to
run sp_dropmergearticle there will be an error if the snapshot has already
been run, so we have to set @forceinvalid_snapshot to 1, make the table
change on the publisher then readd the article and subscriptions and
initialize which would necessitate a new snapshot generation of all articles
in this publication. A nosync initialization is possible, but this can be
extremely restrictive for future changes, and I'll leave that for another
article.'
Does this mean that if I do any simple schema change in merge replication
environment using 'altering subscription' , complete snapshot for all
articles will be transferred over network?
Or is there any other way how to change one article without huge traffic?
Thanks for answer
Lenka