We are trying to build topology for InfoSphere 11.3 installation, to have Repository tier using Oracle database (12c) which will be in Physical Machine (Solaris) while Services & other Engine tiers on Separate Virtual Machines (RHEL).
As IBM recommonds Repository & Services tiers should be in Same Subnet. In addition from your experience, can you please lets know for any recommondations, constraints, and how performance will be if we go by this approach.
Repository in Oracle on Physical Server and Services on VM
Moderators: chulett, rschirm, roy
Remote Repository Installation is supported, but one needs to pre-create the Metadata Repository and (if needed) Information Analyzer databases using the scripts in the installation media. I believe that the only database schemas to be precreated are XMETA, XMETASR and (if needed) IAUSER.
See Pre-installing the metadata repository database in an Oracle database system link. Also see:
http://www-01.ibm.com/support/knowledge ... lldbs.html
As for question about performance/latency ...
The metadata repository holds information pertaining to the jobs, as well as operational and business metadata that you may input to the system.
Interaction with Information Server is first routed to the system holding the WebSphere Application Server component, which in turns sends the request to the metadata repository.
Frequent interaction between the App Server and the metadata repository is an important consideration for having these components on either the same box (same as IBM recommended), or at least on boxes that are connected with very high speed connection. If both of the above mentioned components are not on same box, latency may manifest itself in access from the clients, and in jobs, it will be at the time of startup as well as any operational metadata that you may be keeping.
See Pre-installing the metadata repository database in an Oracle database system link. Also see:
http://www-01.ibm.com/support/knowledge ... lldbs.html
As for question about performance/latency ...
The metadata repository holds information pertaining to the jobs, as well as operational and business metadata that you may input to the system.
Interaction with Information Server is first routed to the system holding the WebSphere Application Server component, which in turns sends the request to the metadata repository.
Frequent interaction between the App Server and the metadata repository is an important consideration for having these components on either the same box (same as IBM recommended), or at least on boxes that are connected with very high speed connection. If both of the above mentioned components are not on same box, latency may manifest itself in access from the clients, and in jobs, it will be at the time of startup as well as any operational metadata that you may be keeping.