Local and Cloud Backup Strategy

[LThibx]

A client of mine has a single Hyper-V Win 2012 R2 Essentials (ESS) VM that I am currently backing up with Veeam (v7) to a NAS shared folder, and all is well with that backup. Win ESS has the capability of backing up the workstations, and I am utilizing this feature. These workstation backups occupies 66%+ of the used space on the VM disk volume. Currently all OS and data is on one VM disk volume. Being a small client (1 VM, 4 workstations), the client is concerned about keeping costs down in getting this data to the cloud, so I am looking for ways to reduce the ongoing size of the backup. My idea to do this is to move the workstation backups to a separate volume (from Cvol to Dvol), and then exclude that DVol from the Veaam backup. To do this I would create a separate backup job saving to a separate repository (separate shared folder on same NAS), and I would back up that repository folder to the cloud. Thus I would have one job backing up the whole VM to repository 1 (which would remain local), and a second job to backup up the VM again, excluding the workstation backup volume, to repository 2. I will then backup that to the cloud, with a Backup Copy Job. Also, I would like to use a Backup Copy Job to retain about a years worth of history on local NAS storage (using repo 1 as source)
.
To illustrate:

VM Size: 250GB / 153GB Used.
Workstation backup size = 105GB
NAS Capacity Free = 4TB+
Currently 14 days worth of backups occupies approx 340GB of NAS storage. Most of the change rate coming from the continuous workstation backups.

Current:
Backup Job 1 (Incremental, Synthetic Fulls -Weekly, Transform, RPs = 14) ---> Complete VM (CVol, DVol-Workstations = 153GB) ---> Repo1 (NAS - Local backup)
Proposed (Add Backup Job)
Backup Job 2 (Incremental, RPs = 14) ---> VM (exlude DVol-Workstation Backup = Full backup size would be 48GB) - Repo2 (NAS Local)

Proposed (Add 2 Backup Copy Jobs)
Backup Copy Job 1 (Retain = 7, W=4, M=12, Q=0, Y=0) ---> Source = Backup Job 1 (Long term storage).
Backup Copy Job 2 (Retain = 14) ---> Source = Backup Job 2 (Keep 2 weeks in the cloud).

Questions
Am I overly complicating the process? Is there a simpler / better way?
If this is a valid plan, would I need weekly Synthetic Full on Backup Job 2?
P.S. I know it would be slower than a Veeam local to Veeam WAN setup, but can I point Backup Copy Job 2 to an Amazon S3 bucket (or similar) without additional software?
This would eliminate me having to have compute in the cloud at this time. (waiting on V8 Cloud Connect to come out, then I would be glad to setup cloud compute resources, as I will become cloud host for multiple clients).

Thanks in advance.
LThibx

[veremin]

Am I getting this right - you're backing up workstations to a volume inside a VM and then backup that whole VM to NAS device? Wouldn't it better to point those backups directly NAS, so that, restore procedure wouldn't be overcomplicated? As to sending backups to cloud, I'd recommend waiting for v8 release and utilizing Cloud Connect feature. Thanks.

[LThibx]

v.Eremin
Thanks for stirring up the thought process. This morning, I created an iSCSI target on my NAS (as it requires an NTFS volume), and changed my Client Computer Backup properties to move that folder to that drive. That will move 105 GB out of the VM, and then will only be an ongoing local backup for those machnes,which Win 2012 ESS will handle. This could be a good situation to handle with Veeam Endpoint when it's released, but for now this will work.

Based on the above, should I see a vast decrease in size on my next Synthetic Full or will I have to execute a new Active full? Should I archive the current backup chain, and start fresh?

Now I will only use the current Backup Job, and create the two Backup Copy Jobs as stated before, one to create long term archival to the NAS local storage, and the second to send to the cloud:

Current:
Backup Job 1 (Incremental, Synthetic Fulls -Weekly, Transform, RPs = 14) ---> Complete VM (CVol) ---> Repo1 (NAS - Local backup)

Proposed (Add 2 Backup Copy Jobs)
Backup Copy Job 1 (Retain = 7, W=4, M=12, Q=0, Y=0) ---> Source = Backup Job 1 (Long term storage - NAS).
Backup Copy Job 2 (Retain = 14) ---> Source = Backup Job 2 (Keep 2 weeks in the cloud).

Does the proposed Backup Copy Jobs seem correct for my intended purpose?

I can't really wait for V8 and Cloud Connect, as my customer doesn't feel fully protected and would like to get this out to the cloud. So I hope to implement a temporary process using storage from Amazon, Google, or similar. My backup job should now see a lot less change rate and will aid in WAN transfer. When V8 is finally released I will switch over to it.

Any additional thoughts and opinions you might have are greatly appreciated.
Thanks for you input.
LThibx

[veremin]

Based on the above, should I see a vast decrease in size on my next Synthetic Full or will I have to execute a new Active full? Should I archive the current backup chain, and start fresh?

Have you removed the corresponding drive completely or just cleaned it?

Does the proposed Backup Copy Jobs seem correct for my intended purpose?

It depends on the type of the cloud you're talking about. The usage of backup copy job will imply that there is a Windows-based machine on the "target" side. If chosen cloud proposes such an option, then, you're good to go.

Thanks.