Re: cdr replication

Thanks. If I remove g= I can't keep server1 in DBSERVERNAME. I get (onstat -m): 13:57:44 CDR no group entry value for server1 -- J.RAGE FRAME SA

Did you try to switch the DBSERVERNAME and alias?? I should have mentioned that ER checks the DBSERVERNAME to ensure that it has a correct ID associated with it. What happens during a connection is that either side of the connection will attempt to connect to the other side. In your case, group1 will attempt to connect to group2. At the same time group2 will attempt to connect to group1. Each side will run through the various paths to the remote server. If one of the paths fails, then it will attempt the next. So one stragety might be to order the connections within the sqlhost file so that either could be used. Since ER will attempt to use the first one before the second one, then that would tend to cause ER to use a different path than the client programs. However, if the first path should be bad, then it would switch to the second. This might provide a failover while also providing normal isolation of the network traffic. Jacques Rage wrote: > Thanks. If I remove g= I can't keep server1 in DBSERVERNAME. I get > (onstat -m): > 13:57:44 CDR no group entry value for server1> > -- > J.RAGE > FRAME SA -- Madison Pruet =========================================== Enterprise Replication Product Developement Dallas, Texas Informix Software ===========================================