#Restore performance for Blue XP Backup?

1 messages · Page 1 of 1 (latest)

little igloo
#

Has anyone actually benchmarked the real world restore performance of BlueXP backup from azure blob to Azure Cloud Volumes ontap? Have a customer with a 32TB (mostly marketing and cad files - mid size) who is looking to potentially use it in that fashion and as an at once 24-48HR RTO instead of paying for the replica copy and CVO licensing constantly.

amber socket
#

Despite other factors it heavily depends on which tier you're using for your Azure blobs. If the backups are stored in Archive access tier then the blobs are offline and first need to get rehydrated. Depending on your retrieval priority you need to wait between 1 and 15hours before it even starts the restore process.

#

And even if you choose the high priority (which is also not guaranteed) only a "maximum of 10 GiB per storage account may be rehydrated per hour".

#

So if you want to use BlueXP B&R for long-term archival but still need a "fast" restore-method in an desaster-event either make sure the first days/weeks of your backups are in an online-tier (only "Cool" is currently supported) or only restore the most necessary files first. Usually you would not need all 32TB. You could use single-folder restore to focus on the most important files and/or work with indexing.

little igloo
# amber socket So if you want to use BlueXP B&R for long-term archival but still need a "fast" ...

While I get all of this math... Cool would be the tier we send it to (too bad wasabi gets snubbed) and in my travels cool and hot are similar perfomance just with different billings rules... I just need a real world TB/hr restore time... Doesn't seem like it would be too much to ask and that someone at NetApp would have done this before selling a product but it is NOT documented ANYWHERE! And honestly the 32TB is the file share that is the most used 😦

amber socket
#

I was not talking about the restore target but the source. If the source backup of your restore currently is in Azure Archive tier then it will be slow. If the backups are in Cool tier already then it's fine.
Regarding restore-performance: Sorry can't help you with, we only tested it with restoring to an on-prem system.

amber socket
#

I guess look for any performance related doc for CVO since a BlueXP restore would simply be a reverse SnapMirror-Cloud with source Azure-bucket and destination IC-LIFs of your CVO-node. There are some TR if I remember correctly.

#

I would ask your NetApp sales team, they should be able to reach out to the BlueXP TMEs.

little igloo
tidal plover
#

You could test on a smaller dataset too. I have no idea. Azure's performance may be a differentiating factor as it may not be consistent.