System Recovery 2011 image replication for off-siting

[Lance_Williams]

Hello all

I have come across the following issue and I hope some of you may have seen the same issue and have a workaround. At Vision in Barcelona this week, some of the experts alluded to breaking the SSR snapshot down into smaller chunks via some additional functionality made accessible by pressing and holding 'CTRL' on the management console.

Here's the scenario:

Servers run SSR locally and write snapshots to volumes presented on the live SAN, and a remote snapshot of this volume is taken to the DR SAN every night.  The problem is that there is nothing clever in the way of deduplication or block level replication going on here, so if 100GB is written to the local SAN that same 100GB has to be replicated to the remote SAN overnight which obviously isn’t feasible across a 2MB WAN link. 

We can minimise the amount of data being written to this SAN by avoiding snapshots of data volumes for example where possible, but I do have to have snapshots of all my servers for this to have any value so there is a limit to how much this can be reduced.  You also suggested that we only take an initial base snapshot of each recovery set and then take incremental snapshots thereafter but I’ve tried this and you have to recreate a new base snapshot periodically.  For the time being I have extended this to the maximum possible so that a new base snapshot is only created annually to see how that affects the daily transfer rate, but this isn’t a real solution as we are just delaying the problem.

Thanks

[criley]

Lance,

You can use the 'Divide into smaller files to simplify archiving' option which simply means you have x number of small (i.e. 1GB) recovery point files instead of one large file.

This does not change the total amount of data that needs transferring across the WAN but it might help if the issue is with copying/moving a single large file.

Depuplication is on the roadmap but I cannot give you an ETA I'm afraid.

Let us know if you have any further questions.

[Lance_Williams]

Hi Chris

Thanks for this; we'll test it out today and see if it helps the situation.

Personally we'd like to see the line upgraded from 2MB to 10MB at least, but it's not in the budget!

[mistong]

Hi I am looking for some information on setting up our backup like suggested in this post.

We are looking to make a base snapshot of our current system onto a server. We then want to move this server to an offsite location with VPN access.

We then want to make incremental updates to this image.

I was wondering is there a white paper for this or some other information?

[criley]

No white paper exists for this scenario as far as I know.

However, the procedure is pretty simple. As long as you can access the offsite location from the source server then everything should work fine. The only potential issue you may have is moving the initial full recovery point. If this is several hundred GB then you may have some bandwidth issues. All depends on the infrastructure you have in place..

[mistong]

If I save the backup image to the server locally over the LAN.

Then move the server off site and do the update images over VPN.

Would this work OK?

[criley]

Offsite copy just does a standard file system copy of the recovery points. As long as you have no network issues, it should work fine.