Feature request: FLR restore from remote site Data Domain

[rasmusan]

Hi

I recently ran into an issue at a customer, which after opening a support ticket appeared to be by design.

To sum up briefly:

Customer is running VBR v9 Enterprise.
Has a primary site, where Veeam backup server and repository (JBOD) for Backup jobs are located.
They also have a secondary site, connected via 100Mbit/s WAN, where repository (EMC Data Domain (DDBoost)) for Backup Copy Jobs is placed.
Gateway server for Data Domain appliance is the VBR server in primary site, to leverange DDBoost over WAN (source side dedupe), which works great.
a server placed on the secondary site is designated as mount server for Data Domain

I thought the mount server was responsible for everything including mounting restore point when performing FLR restore - now this turns out not to be the case. The client performing the FLR restore, will mount the restore point. That means, that when FLR restore is launched from the primary site, the FLR mount is very slow - the larger the VM the longer time to mount. An example is their file server, which is around 1,2TB in size - FLR mount times out after about 1 hour.

Now if VBR console is then installed on the server in the secondary site, and the same FLR restore is initiated, the traffic still traverses the WAN, due to VBR server being gateway server for Data Domain.

This means, that to have FLR restores perform reasonable, they have to manually change gateway server for Data Domain to the server in the secondary site AND launch the FLR restore from the server in the secondary site.

Sorry for the long explanation, but I hope you get the point and understand the issue - It would be awesome if the mount server somehow was able to "proxy" the FLR mount, so the mount would not have to go over the WAN

[foggy]

It is possible in case the mount server is the gateway server as well, since it is the only server that is able to talk to the repository and retrieve data (has DD Boost libraries installed).

[rasmusan]

sure, but to leverage DDBoost over WAN, which was introduced in v9 and by the way works great and saves a lot of bandwidth, you have to use a server on the "primary" site as gateway for the DDBoost device...

and either way, you still have to run the VBR console on a machine on the remote site, as the FLR mount always happens on the VBR console where the restore is initiated...

[foggy]

Correct. To keep restore traffic local, you can temporarily (for the time of restore) change the gateway server in repository settings to the one located in the same site.

[rasmusan]

sure, but that seems a little "complicated" - it would be nice if that intelligence was somehow built into Veeam, thus the reason for my feature request... can you bring this to the right people within Veeam ?

[foggy]

"Right" people are closely monitoring these forums, so be sure your suggestion is not overlooked. Thanks for the feedback.