We have implemented clustering at the WEB Tier Layer, Intelligence Layer and Processing layer. Hence, we have the below four machines in our architecture.
A) W1 -> WEB Application Server Machine 1 - 2 Tomcats
B) W2 -> WEB Application Server Machine 2 - 2 Tomcats
C) P1 -> Processing & Intelligence Tier Machine 1 - Here, the directory /u01/app/BO/bobje/data holds frsinput and frsouput directories that comprise the physical repositories for input and output repositories respectively. This also has our Database repositories (for CMS and Audit databases). Also, the IFRS service and OFRS service along with CMS and other services are available here as well.
D) P2 -> Processing & intelligence Tier Machine 2 - Here, we have all our services available (as a part of the clustering of processing and intelligence tier) but no physical file and database repositories. The FRS & CMS services in this machine point to physical repositories located in machien P1 as should be the norm.
However, we recently faced an issue where our machine P1 crashed. The entire BOE system, hence, was unable to endusers (since the Data Repositories & the physical file repositories were unavailable to end users).
We hence plan to build a complete failover where we implement two physical file repositories and two Data Repositories (for CMS and Audit databases).
The problem is how do we synch between the physical repositories. The failover repository should contain an exact replica of the main repository. How do I achieve this.
Any help would be appreciated.
Thanks & Regards,