I faced one issue regarding RFC connection between SAP and webMethod which affected some of the partner transactions.
Our set up is that we have 2 RI server and 2 IS 6.5 in cluster.
One of our IS server is not able to make RFC connection with SAP but other server was up and running. So if both the servers are in clusters then why the request didn’t went through the other one.
Though you have two servers in cluster the RFC Connection object of that IS will be used which is processing the request.Hence any request being processed by the Impacted IS will fail.
I can understand what you are saying and agree with you but one question still remains that our 16 transaction failed at that time so if one IS was down then after few(or 1) transaction failure the request should automatically go through another one.
I am not sure why it didn’t happened. Can you throw some light on this.
have you checked repository logs for second IS to make sure it has taken over once First IS went down? Did other other transaction got routed to it after?