Need help with Deduplication
Noob here. Two things. First how do I setup deduplication. I've read the help/best practices articles and can't figure it out. Can I setup a dedup vault using personal vaults or do I have use centralized vaults? If i need centralized vaults, how do I create one? All the commands are greyed out under centralized vaults.
Here's my setup:
Win SBS 2011 Essentials Server with a dedicated HDD for backup. Offsite NAS accessible through a site to site VPN (2 Mbps link). Due to the slow link, I am looking for the most efficient way to setup my offsite backup. I'm thinking of an annual Full Backup with monthly (or weekly) Differentials and daily Incrementals. I'd also like to use data deduplication minimize bandwidth utilization and to a lesser extent minimize storage utilization.
Can/should I do this with a local backup that is replicated to the remote site or should I have separate local and remote backups? Where should the dedup vaults go? I believe they need to be on the source machine (to minimize network traffic). If the dedup vault is local, will it be replicated with the other data to the remote site?

- Log in to post comments

Thanks for the reply. Maybe you can help me understand some backup theory/best practices.
My main issue is the slow link (2 Mbps although future needs will probably push it down to around 1 Mbps) which makes a full offsite backup take over a week. I understand that incremental will backup only the changes made since the last backup (of any type) which is what I want in my situation. I was hoping, however, to use a dedup vault to "seed" the backup remotely. My theory was to carry a hard drive to the remote site and copy it to the dedup vault. Any subsequent full backups should (if I'm understanding this right) carry minimal changes/bandwidth usage making a monthly full backup more feasible. Is it possible for me to use a dedup vault in this manner?
What is the downside to doing annual full backups? From what I've read online, there seems to be an aversion to having long intervals between full backups but I'm not sure why. A monthly or weekly differential should keep the system from having to reassemble too many backups in a restore (I'm guessing this is the reason to avoid long full backup intervals).
Can you point me to any resources to further my understanding of backup practices?
- Log in to post comments

Deduplication works after the data to be placed in the incremental backup is calculated. If you have only one machine and plan to perform only incrementals and differentials after the first full, there is no use of deduplication. The approach with 'seeding' full backup to offsite will work, but you can do it with a simple personal vault. Just set up the the backup plan with the second vault on some local storage, perform the first full and replication, copy it offsite then change the second vault in plan settings.
Longs intervals between fulls mean that if the full backup is corrupted, it's useless and all its differentials/incrementals are too and you have to use the previous full (if there is one at all). In your case, however, you have a second offsite copy.
- Log in to post comments

Ok. Is there an issue if I use a 3rd party program to handle the replication of the local backups to the remote site? Do I just need to copy the contents of the vault (.TIB file(s), .lck, Catalog folder, and the XML document)? Do I need to copy to another vault or is a general folder ok? I guess I don't quite understand the difference between a vault and a normal folder.
Using a 3rd party program will let me get around the issue of Acronis failing to complete other backup jobs while offsite replication occurs.
- Log in to post comments

There is a difference between unmanaged and personal vaults on one side and managed vaults on the other. Unmanaged vault is a folder and backu plans on agents contain the path to this folder. Managed vault is connected to some storage node that is registered in Management Server and can't be used without a storage node. Agents know the Storage Node address and vault name and send data to Storage Node. Copying files from unmanaged vault (.tibs + xmls; catalog can be recreated in the new vault) should be performed when no agents backup up to the vault. To copy an unmanaged vault (and its deduplication database in case of vault wit hdeduplication) the Storage Node service should better be stopped. Recover from an unmanaged vault copied in such way can be done directly. Managed vault needs to be attached to some storage node first.
- Log in to post comments

So I can create an unmanaged/personal vault at the remote site and copy the .tibs/xml files to it (either manually carrying them to the site or through a 3rd party sync/xfer program) and Acronis will catalog it and recover from it?
Would it be better to create a post-task script in Acronis to call my sync program when the Acronis backup to the local vault is done?
- Log in to post comments

For the first question - yes you can, but if you let it catalog this vault, it will be reading it updating catalog database and consume bandwidth. So I suggest to not create vault on the offsite nas, just sync backup data there. Catalog folder is not necessary for recovery, but if you need it in case of onsite copy is damaged, you can bring it back on a physical drive as well.
Replicating in a post-script will work if you make sure it's done before the next backup starts.
- Log in to post comments