Restoring 43TB and Applying 2.2 Billion DB2 Log Records Within 4 Hours, and Proving It on a Live Production System!
Project and Program: Information Management
, SHARE Atlanta 2016
Most IT installations have Recovery Time Objectives (RTOs) in place for business critical applications, but proving that such objectives can be met is not always a given. In particular, for worst case scenarios like recovery to a Point in Time (PIT), which is entirely different challenge than fail over to a secondary site.
Challenges involve defining and validating subsets of data as you may not have time to recover everything, at least not in the first wave. As such you must focus on what is really required to support truly business critical functions to operate. Secondly you may have to acquire and deploy new technologies to address the need for speed to meet aggressive RTOs. As you start deploying new technologies you’ll learn what works, but more importantly also identify implied pitfalls and understand how to work around them. Finally, you’ll have to prove that everything works as designed and that RTOs can actually be met, which evidently means simulating such PIT recovery scenarios on a live production system while you are still serving your normal business operating in a 24 by 7 all year round modus.
Through this presentation you'll get an understanding about how this was achieved at UBS in terms of well-defined processes and deploying state of the art technology both on the hardware side as well software side • Managing the RTO Objective – Defining Spec's and Requirements
• The different dimensions of Big and what the system setup looks like
• Mission accomplish: Performing point in time recovery simulation with a RTO of less than 4 hours
• Lessons learned and things to keep in mind
Back to Proceedings File Library