Oracle DBs Consolidation to Real Application Cluster (RAC)
Telecommunication and media company
The client needed to create a typical migration strategy with small downtime, a fast failback option, and performing DB migrations in a timeframe not to exceed more that 2-3 hours. In addition, the client wanted a Linux cluster design and deployment (from network architecture and redundancy, ASM disk configuration to DB back-up/recovery strategy).
CoreValue set up a 4-node Oracle RAC cluster on Linux and 5 Oracle databases (with a size from 400G to 2.5T) working on several Solaris standalone servers, which were to be consolidated on the new Linux cluster.
Based on a stable and fast method for DB migration (Transportable Tablespaces in conjunction with RMAN cross-platform incrementals), CoreValue’s team implemented an effective enterprise reporting solution consisting of the following parts:
- Implementation of redundancy for cluster interconnect, to ensure system protection from interconnect failures.
- Disk subsystem configuration, starting from setup on disk array to ASM-level setup. This advanced disk layout was chosen to ensure database independence from each other, providing higher performance characteristics.
- Implementation of two mutually independent RMAN back-up/recovery strategies: SBT backup to EMC DDBoost with replication to DR site, and RMAN Incrementally Updated DB Image Copy.
- Oracle GoldenGate unidirectional replication in order to provide a fast failback option after DB migration.