Inscrivez-vous ou connectez-vous pour rejoindre votre communauté professionnelle.
Oracle RAC and Oracle ASM very nice and very good technology added to oracle, oracle RAC increase the performance and increase the availability. ASM protect from hard disk failure, and RAC protect from Instance failure. oracle Data Guard ensure the High Availability.
You are correct...on top of that now you can use active data guard and divert all your select queries to your standby database. Also what most banks do is they have online storage mirroring of redo logs and control file at a near DR site.
Together will all above solutions you get zero data loss DR strategy.
The Maximum Availability Architecture (MAA) is Oracle's best practices blueprint. It is based on proven and validated Oracle high availability technologies and recommendations. The goal of MAA is provide architectural recommendations based on your business requirements and then optimize and stabilize the architecture and recommended high availability features with configuration and operational best practices.
the absolute best is having2 separated RACs (2 identical hardware setup with different SANs ... etc)
one RAC is setup as the Production and the other as Active Replica as well as having a "Zero Data Loss Recovery Appliance"
RAC is good if you have one Node failure, you will be able to continue with the other node(s) ... but there is a catch in case of the SAN failure .. you will need the Active Replica
now having an identical hardware for the Replica will save you from the performance issue in case if the rest of the production node(s) dose not cover the needed performance.
with the Zero Data Loss Recovery Appliance you will be able to recover loss in data at any giving time, if a user delete some data, it will be deleted from both the Prod and the Replica.
you have to build a system that suits your need based on how critical you data and your budget
Best is to use Dataguard with RAC using SAN over remote locations only if your company can afford the costs but cannot afford any downtimes.
Otherwise cheapest would be to perform incremental backup restore at daily/wekkly schedules in a standard server at remote place,
Reagrds
Rizwan
