zero downtime SRF deployment
Hi,
As we reconfigure component to point a different SRF, is it good to implement this to replace new SRF without downtime?
As per reconfigure, All siebel servers in the enterprise creates a number of new processes (MinMTServer) after "commit reconfiguration" is triggered.
If an ObjMgr has 5 minMTServer and 5 MaxMTServer and MaxTasks = 100.
Now lets, suppose currently 15-20 users are present on this component and at least 1 user is on every process of this ObjMgr.
Now if I reconfigure that component, will it create 5 more processes for the component with new definitions. And if the old 15-20 users remain active for around 10 hrs, the server would be handling double the processes which it should? So reconfiguration can the number of processes double and hence increasing the load on server. So how to determine, how much server capacity should good for any component reconfiguration?