UNSOLVED
R
Ryan_Johnson
1 Rookie
•
73 Posts
2
2253
September 27th, 2016 11:00
Data Domain Cloud Tier DR
Is there a way to get the data from the Cloud Tier if the data domain is lost? I've read the DDOS 6.0 release notes and administrator guide and I didn't see this scenario mentioned. Under Using DD Replicator is does state that recall of a file from the source cloud tier to the destination is not allowed (it has to be recalled from cloud tier at source then replicated from active tier).
Would the local data domain be a single point of failure for the data in the cloud tier? For example if the Data Domain is lost due to a natural disaster is there anyway to recall any of the files in the Cloud Tier to a different data domain?
Is there any way to do a planned migration of the Cloud Tier from one data domain to another? In particular I'm thinking when hardware needs to be refreshed it is not going to be practical to recall all the cloud tier data to active tier, replicate to the new hardware and then transfer to a new cloud storage unit.
Ryan_Johnson
1 Rookie
•
73 Posts
1111
0
Posted October 14th, 2016 12:00
I was able to asked this question to an EMC data domain project manager that I ran into at an EMC event. My understanding is DR functionality is coming soon. Possibly next release.
So in DDOS 6.0 to protect against this single point of failure you would need 2 data domains each with a cloud tier.
Hopefully in 6.1 there will be a documented and fully supported way to recover the data in the cloud tier in the case of a disaster.
ble1
6 Operator
•
14354 Posts
•
56186 Points
1111
0
Posted September 28th, 2016 12:00
Hmm... good one. From what I understood, cloud tier get's only unique data, but is it just unique to cloud tier and in my view this data should be such that you can recall the file(s). I know that in theory when calling it back it retrieves again only unique data which is cool, but in case you loose original DD I guess some sort of rebuild would be needed. But I still expect that all data on cloud tier is enough to rebuild database and have data returned as otherwise that would silly single point of failure. Perhaps someone from DD group can confirm...