AnsweredAssumed Answered

Disaster Recovery Fail Back

Question asked by dzndrx on Jun 19, 2017
Latest reply on Jun 20, 2017 by dzndrx

Hi, We are currently want to test the DR capabilities of MapR.

 

We just want to validate this scenario, or if this is feasible. 

The machine specs are not yet clear but generally as my assumption it will be like this.

 

DC = Data Center

 

Total be 3 nodes in DC and another 3 node in data center 2

CPU: 16~32 between these values

Memory:128 GB

Disk: 10kRPM disk plus SSD's (around 22TB total cluster capacity)

 

For example, we have setup a active volume on DC1 and schedule a mirror to DC2. In a scenario which DC1 goes down, we still have a copy on DC2 which is a Mirror Volume (Read Only) and if we need to go back to production, we need to promote this Mirror Volume to an Active volume hence solving the most common DR purpose.

 

Then lets continue, since we are now using the active volume of DC2 for a quite time, we managed to recover the DC1 cluster and retrieve some of its data (this is just an example) , now we are ready to make the DC1 as the active volume and use the DC2 as the mirror volume. What is the process on this? Does this means that I need to make a Mirror Volume in DC1 in order to mirror the active volume in DC2, but what if I recover the DC1 and still have the historical data of the active volume before it goes down. Can I just make this active volume on DC1 set to a mirror and start mirroring the active volume in DC2 which results only getting the delta data from it?

 

Does this make sense to you guys? Correct me If Im wrong. I really need an assistance for this.

Outcomes