Unanswered: sysmergesubscriptions table in Merge process
I am working to develop an application that will use SQL merge replication to move data between databases. The application will be creating and deleting a lot of merge subscriptions on a single publication. I am having problems with entries in the sysmergesubscriptions table on the publisher not getting cleaned up. There are a lot of old entries in this table. These old entries are causing conflicts when I try to create a new subscription on a system that has had a previous subscription to the publication. I am looking for a process that will clean up these old and deleted subscriptions from this table.
Thanks for the reply, however, the suggestion does not resolve my issue. My application will be using SQL Server to store configuration details that are recorded by our application. Many engineers will be doing design work that will be stored within the database. I am planning to use merge replacation with dynamic filters to distribute portions of the data to different servers so I can limit the total amount of data at the different locations. The design uses several layers of publishers and re-publishers to distribute data to multiple design centers. The intent is to allow application users the ability to connect to a publication and create a subscription that will replicate small portions of the total database to their local system. Users will be able to add and drop subscriptions as the project work evolves over time.
Part of the design requires using attachable subscriptions. The attachable subscription database will be loaded on a system that does not have access to the original publisher. The intent is to use the alternate synch partner feature to allow replicating data to these disconnected servers.
During my design work to support creating all of these subscriptions and attachable subscriptions, the sysmergesubscriptions table gets loaded with entries that prevent me from attaching subscription databases to the same server more than once. I am trying to identify either a manual or automated process that will clear out these previous entries. My current sollution is to totally remove all replication support on four systems and recreate the publisher-subscriber chain through all of these systems. (This is a somewhat painfull process.)