Temporarily stopping replication on a server

You can stop Enterprise Replication on a server by shutting down the database server. Replication begins again when you restart the database server.

However, you might want to temporarily stop the Enterprise Replication threads without stopping the database server.

You can temporarily stop replication by running the cdr stop command. The stopped server does not capture data to be replicated. Other replication servers in the domain continue to queue replicated data for the stopped server in their send queues. Replication threads remain stopped (even if the database server is stopped and restarted) until you run the cdr start command. When you restart replication on the server, it receives and applies the replicated data from the other replication servers. However, if replication is stopped for long enough, the replay position on the logical log on the stopped server can be overrun and the send queues on the active replication servers can fill up. If either of these situations happens, you must synchronize the server that was stopped.

If your replicates use time stamp or delete wins conflict resolution rules, you should temporarily stop replication on the server by using the cdr disable server command. Disabling a replication server is also appropriate if you do not have enough disk space to avoid overrunning the replay position. Replication servers do not queue replicated transactions for the disabled replication server, nor does the disabled replication server queue its transactions. Therefore, you must synchronize the replication server that was disabled after you enable replication on it by using the cdr check replicateset command. However, because information about deleted rows on the disabled replication server is saved in delete tables, you can take advantage of a time stamp repair.