Wan Accelerator usage

[ksims]

Alright! Lets talk WAN Acceleration.

Currently, I'm backing up ~5 sites using Veeam, and have been trying to get their offsites going as well. I've created seed copies of the data, brought that back to our offsite replication location.. Getting it moved over to the offsite storage is all simple enough - The issue I'm running in to is that most of the offsite copy jobs are running at ~500kbps. I'm assuming that if I set up a WAN accelerator at each client location and one at our offsite storage, things will speed up.

The questions I have are these -

1. What type of load does the WAN accelerator process place on a workstation? I think more specifically, is it viable to run in the background on a workstation or will it affect machine performance too much?
2. Will I need multiple machines at the offsite location, or will one WAN accelerator job be enough?
3. If I need more accelerator 'jobs', can you set up multiple on one machine?
4. What type of offsite copy jobs are people doing? Incremental, or periodic? I keep telling myself that I want to do incremental offsite copy jobs, but given my current situation, this just doesn't seem feasible... The copy jobs aren't able to catch up, because they're taking ~5 hours to copy a file, and my local backups run every hour. The backlog just keeps growing.

[Gostev]

1. It should be viable given your primary constraint seems to be WAN bandwidth and thus WAN accelerators won't need to sweat much.
2. You only need one WAN accelerator machine per site.
3. Sure, you can have multiple jobs using the same WAN accelerator.
4. WAN accelerators particularly shine at accelerating incremental backups and I would definitely expect reduction of transfer speed above 5x once you have them deployed, meaning you should be be able to do the Immediate Copy mode (as copies will be taking under 1 hour).

[ksims]

A new related question/issue -

I realized that part of the problem I've been having with the WAN accelerators, is that I wasn't pre-loading the cache on my source accelerator. To handle this, here are the steps I followed -

Logged into my Cloud Connect server and loaded VBR (Licensed for Cloud Connect)
Rescanned the repository
At this point, it says it imported some backups, but when I go to the Home view, there are no backups showing. I tried to manually import a backup, and it says it already exists.
My backups do use encryption, so I know I need to use my password to unencrypt them.. But I can't get them to show in in the VBR console on my Cloud Connect, in order to do that.

[Gostev]

For any technical issues, please open a support case as we can't troubleshoot them efficiently over forum posts. Thanks!

[veremin]

Aren't the backups you are trying to import the backups created by your tenants by any chance? Thanks!

[ksims]

Correct - The backups are created by tenants.

So, my 'Seed process' looks like this -

Create a Backup copy job on tenant server
Backup copy to external HDD
Bring HDD to our data center
Copy data to our repository
Rescan cloud provider repository (Our datacenter) from client VBR
The backup files show in the client VBR, and are already unencrypted - Assuming because thats the same machine that originally encrypted them
Backup files do not show in our Cloud Connect VBR, and we cannot cache our destination WAN accelerator, because they're encrypted, and I can't unencrypt them because of the above issues

[veremin]

Be aware of the limitations mentioned here:

Veeam Backup & Replication does not use encrypted backups for global cache population.

[Veeam Cloud Connect] Veeam Backup & Replication does not use tenant backups to populate global cache on the service provider side.

Thanks!

[ksims]

So.. How do you populate the cache on the service provider side, if not with tenant backups?

[Gostev]

In a typical Veeam Cloud Connect case, tenant backups are not an option in principle, because they are encrypted at source specifically to prevent service provider from retrieving data from them. However, you could create your own backups of VMs with OS your tenants use, and use those backups to populate the cache. The only goal here is to capture blocks belonging to OS and standard apps modules, as this will be the only thing common in data coming from different tenants.