Seed virtual machine in backup job

[SebastiaanT]

I have a bacup job containing some virtual machines. These are located on different remote sites and each remote site has it's own backup proxy. All remote sites have a connection back to a central backup location but upload speed is limited. As a result, every time a new virtual machine is added, the initial copy can take a week (when limited to 3Mbps) meaning that during this period new backups of the other virtual machines in the job are not occurring.
Is there a method of seeding the first backup of a virtual machine by using a USB disk connected to the backup proxy, transporting it back to the central location and importing it in the backup job?
Storage on the remote site is limited so using a local backup job and replication job (where seeding is supported) back to the central location is not an option as there is not enough local storage to support that.
Any help would be welcome.

[foggy]

Hi, there's no easy way to do this. Theoretically, provided per-VM chains are enabled on the target repository, you could create a full backup of this VM and put it into the remote repository, however it will not be picked up automatically without editing VBM file, which is not trivial/recommended.

It is always recommended to perform local backups and then copy them offsite using WAN acceleration for slow links. In this case, you would have the option to re-seed the job in such situation.

[jmmarton]

Would it be possible to do this?

- create a backup repository on the proxy at the remote site pointing to a locally attached USB hard drive
- perform a backup
- ship the USB drive to the main site
- import the backup
- map the backup file to the job
- edit the job to use a repository at the remote site

The end result would then be that incrementals would go directly to the remote site and the backup chain would be intact from the imported & mapped full backup file? All of this assumes per-VM backup files.

Joe

[foggy]

Joe, do you mean a separate job for the added VM? I was more thinking of adding it into existing one.

[jmmarton]

I didn't think about that. If it's done with the existing backup job, then you can't specify one repository for one set of VMs and a different repository for the new VM. Yeah it would probably have to be a new job for the new VM.

Joe

[foggy]

Right, and my impression was that Joe was asking about seeding the VM into existing job.

[SebastiaanT]

@Joe, I did try this but then the existing chain is discarded.
@Alexander, I am using per-VM chains and I allready did the usb backup locally and did copy it to the repository but as you mentioned, the job doesn't pick it up and starts a full backup.
Can you give some info on how the vbm file should be altered to pick it up?
FYI, I also have a case open for this.

[SebastiaanT]

I understand the preferred method is a local copy and using a copy job with WAN accelerator to replicate. This would mean only 50% of the local storage can be used and this is not the case.
Another challenge with this is that we are keeping two copies stored on different storage devices by using two copy jobs. In this case, the advantage of less data over the WAN is lost because it has to send everything twice over the line as a copy job from another copy job is not possible.

[foggy]

Can you give some info on how the vbm file should be altered to pick it up?
FYI, I also have a case open for this.

I believe our support team can assist you with this.

as a copy job from another copy job is not possible.

Copy job from another copy job is actually possible, no issues with this.

[SebastiaanT]

In a copy job, I only have the option to select machines from infrastructure, backups or jobs. In the last two, I can only select backup jobs as backup copy jobs are not listed.

I updated the call regarding a procedure on how to change the vbm.

[foggy]

You should be able to select backups created by the backup copy jobs as well (using 'From backups' option) and limit the scope by specific repositories. Using the 'From infrastructure' option is also possible.

[SebastiaanT]

Thank you Alexander, I will definitely give it a try.
If I have a procedure to change the .vbm file, I'm all set

[SebastiaanT]

Support contacted me and told me it's not possible to do this
Could anybody help me on what needs to be changed in the vbm file?

[foggy]

Have you tried to select the backups created by the first backup copy job as source for the second? As I mentioned earlier, editing VBM file is not recommended.

[SebastiaanT]

I tried this but in "from backups" the backup copy job is not mentioned.
In either way, this is not a solution as there is not enough storage on the remote locations to support one local copy.
I can understand editing the VBM file is not recommended at this point but I hope you can see the requirement to do so.
Another case could be when you reallocate a vm from one job to the other and don't want the full backup to run again.
After all, the vbm file just contains the pointers to the data files which I can copy over.

[foggy]

Looks like a feature request.

[SebastiaanT]

That's exactly what support told me