Recently I had received a hand-off ticket which was about a replication issue. The system has been configured with the replication of Peer-to-Peer type.
One of the subscribers was throwing an error which was reading like A conflict of type ‘Update-Update’ was detected at peer 4 between peer 100 (incoming), transaction id 0x000000000011a1c3 and peer 100 (on disk). While I was working on this issue and trying to resolve it, I noticed that it wasn’t showing any records in msrepl_errors table or conflict_dbo_table.
Here again, the Error Logs help, as they have the complete details logged in, which help us identify the table name and exact error with the record. If that hadn’t been the case, I would have followed the Replication Troubleshooting method describe in KB 3066750 to fix the issue.
At this time, with the information we had in hand, we reached out to the customer and resolved the issue by fixing it manually. I would like to mention that there are always two ways conflicts are handled in P2P replication:
1) Manually fix the conflict/data issue
2) Let the winner node have precedence about the data/conflict In P2P replication
At the time of configuration, we will have an option to choose which node will have precedence and can be declared the winner. This is decided by the way of originator_i; the highest originator_id will win. We will have to decide this carefully, as once the setup is done, orginator_id is allotted it can not be altered later.
Here are few reference article that will help you understand this topic better: