Off Site Backup and Replication

[zdundore]

I am currently trying to get a client of mine setup as our first off site backup user. I have successfully created a backup on my USB HD and have transferred the data to our Veeam server on our site. I now would like to configure their backup job running on their server to backup to our server so that the backups can replicate with the existing files on our server. Is there a good technical resource for how to get this done?

[Gostev]

Hi Zach, I am not sure I understand what you are trying to do. Are you trying to seed full backup file?

[zdundore]

I have already seeded it. I ran a backup at my client's site to my external HD on Tuesday. I took the HD to my server offsite on Wed and copied over the backup files to our Veeam backup server. I have our firewall configured to route and allow Veeam traffic to the Veeam backup server, so now i would like my client's Veeam Backup software to start backing up to those backup files i have copied to our backup server. Is there a technical doc or website that could help with this setup?

[Gostev]

OK, basically it is impossible to do what you are thinking (backing up from one Veeam install to another Veeam install). Not even sure where you've got this idea from, this was never possible...

This is what will work:
1. Setup a Linux server on your side (minimal install, Perl and SSH are only components required). This is where you will store your customer's backups. Minimal VM config should be sufficient (256-512MB RAM with console only install).
2. Setup Veeam Backup at your customer's site. Customer will most likely want to backup all VMs locally anyway (for fast operation restores). And subset of those VMs (most important), customer will want to backup remotely as well (to have an offsite copy).
3. Add remote Linux server set up during step (1) to customer's Veeam Backup install via Add Server wizard.
4. Setup a "remote" job on customer side with those VMs customers wants to backup offsite. Specify local destination for now (local disk). Run the first job pass, make sure its successful. Disable the job temporarily (right-click the job > Disabled).
5. Transfer backup file to your Linux server on USB hard drive, put them into required backup directory.
6. Change backup destination of "remote" job on your customer's side to a Linux server added in step (3). Remember to specify the path to folder where you put your backup files in step (5).
7. Have the customer enable the "remote" job, and continue running it normally.

Thanks.

[zdundore]

Ok, i understand what you are saying and i guess i am not wording this right. I have SAN location attached to our Veeam backup server on our site. I am not really trying to back it up from one Veeam install to another Veeam install, i am just trying to access the SAN storage location attached to our Veeam Server (Windows 2008 R2) Can i access that storage in windows as a valid backup location from my client site using Veeam? Or does it have to be a linux box?

[Gostev]

Theoretically, you can do this (make a regular windows share on your side, and have customer specify it as a backup target). In practice, performance of this solution will be unreasonable over WAN. On the other hand, using Linux server allows Veeam to deploy an agent on your side that will handle creating full backups from incremental data (forever incremental). This reduces bandwidth requirements dramatically.

We are planning to enable the same technology for Windows servers as well in the next release, but for now, it has to be Linux server.

[phoerter]

Anton,

So when using a linux destination over WAN, when the synthetic full is created performance is much better than a Windows share destination? Is the data transmitted over an SSH tunnel? Do you have more details in another thread?

Great discussion!

-Patrick

[Gostev]

Correct, performance will be tons better because data processing will be handled locally (versus over WAN as in case of Windows share).

No, the data is not transmitted over SSH (that would be too slow)... just regular socket to socket connection, port 2500 (and the following ports, in case of parallel jobs). SSH is only used as a transmission control channel.

Sure, there are plenty of existing topic discussing using Linux as target, many customers are doing this for remote site backup (it is really the only good option now), so just search the forum... although I've already said above everything you need to know to proceed

[phoerter]

Lurker?! Awww, can't I be upgraded to troll??
Okay okay, I'll contribute more..

Thanks for the response. Sorry for hijacking your post Zach, I owe you a cold beverage of your choosing.