How to manaully Sync the subcriber data in Disaster

First, run tablediff utility to find the difference in data and copy this difference to a table (this is for backup purpose if anything went wrong). The option is to reinitialize your subscription, here you have the options Upload unsynchronized changes from the subscribers before reinitialization. Once the changes are uploaded the merge agent will overwrite the subscriber data with the punlisher, This applies only for merge replication.

..script publisher\subscriber (later you need it to recreate publication)
1. have a good backup of publisher and subscriber DBs
2. drop subscription and publication
3. update publisher from the subscriber backup (please clarify -... if it is table copy, db restore , etc)-- if db restore you mayneed to do some extra steps  -
4.make sure your  "to be again" publisher db \tables look as you expect
5. recreate publication\subscription(s)

------------------
BTW:  what repilcation are you using ? is it snapshot publication?

I am using merge replciation and the steps you have suggested is not possible as the data is updated at both the ends

regards
shailesh

###

One thing I have found over the years is that what users ask for isn't usually a good thing to give them if I have not checked out the full scope of the effect.  One of the reasons i like the table diffs is that it will give you a heads up as to whether any other changes are going to be made other than what you are expecting.  I've aften been requested to do a full restore only to discover that the request was made totally ignoring another oeprational area that would have been adversely impacted if I have done a full restore.

I am also not sure how conflict resolver plays in here.  That only identifies conflicts Sql Server cannot resolve if two changes are made to the same entity at the same time and Sql Server could not determine who should win. (Unless I am missing a capability that I am unaware of and would be extremely interested in exploring further...)

Solutions4Problems:
please read EE help about C grade -- it is like to say somethin bad to people who try to help you
So not nice  :(

if you got it - try to change your grade

Perhaps it is our obsession about metrics that needs to change :-)   Sometimes we will give a user exactly what they want and sometimes we won't.  The is that old saying that if you are not making mistakes, you aren't doing anything.  Or at least, you are not trying to expand your horizons.

So what is wrong if we do not always get an A ?  Why should we be evaluated only on a grade rather than the content of what we said or the fact that we tried to help.

But then I stopped participating in peer code reviews whan management decided they needed to be 'measured' for 'flaws'.  That immediately ended the free flow discussion of what might have been done or techniques that could be sued the next time. Instead, it became a tool for scoring points against those you disliked and adding points for those you did.  All by whether or not you listed the 'flaws' you found...  But management , in particular, like easy answers they can poinit to for the illusion of quality control and objectivity.  I have not met an 'objective' measurement yet that could not be corrupted once people knew it was being measured...

Sorry for the long reply.  I wish people could acccept that grades are relative to peoples expectations and sometimes their expectations can't be achieved as they want...