The proper way to seed a backup

[fafa24]

Dear All,

I would like to reassure myself, if I do the proper way to seed a backup for offsite

1. I create a backup copy job on the source server to a repository on the source server
2. The copy job will create a vbm and vbk file in a folder called after the backup copy Job Name
3. I will copy the folder under 2 on a USB drive
4. I will copy the folder on the USB drive to the target repository
4. I will do a re-scan of the target repository on the Veeam console
5. I will amend the Job (created under step 1) to point to the target respository and map the backup file

Question:
1. Is this procedure correct?
2. Do I always need to map?
3. What is if there is already a incremental file in the backup chain. Is this a problem?
4. Is there a way to speed up the WAN cache? I was reading that I could populate the cache. Is this something I will need to do in the source WAN?

Sorry for all these questions. I'm trying to be sure to do it most efficient to save WAN bandwidth and time.

Thank you!

[Shestakov]

Hello!
1. You need to create a backup job, not backup copy first to get .vbk, then you copy it offsite (using USB), rescan it there and map just as you described. The procedure steps can be found here.
2. No, just once, then you will have incrementals in the target-site chain
3. You can copy and rescan Full with its incrementals as well
4. Yes, you can populate global cache. You need to go through 4 easy steps described in the user guide.

Thanks!

[Vitaliy S.]

You can also use backup copy job targeted to a local repository, run at least one successful pass and then move the entire repository to the offsite location. In this case you don't need to do seeding/mapping, just keep running your backup copy job targeted to the repository you've just moved offsite.

[Shestakov]

But keep in mind that you can`t use same repository as a target for backup and backup copy job, because there is built-in mechanism preventing backup copy job from looking restore points in the repository that is specified as backup copy job target.
Thanks!

[fafa24]

Thanks for all your replies.

I made a backup Job and then put on a USB drive and move it to the target repository offsite, rescan repostory and mapped. So far all fine, but it takes a long time to complete the first pass of the backup copy job. The Job contains 9 VMs, total 2.2TB. The backup window during the week is from 10pm to 7am. It runs each night and it's working, but slow. It did that for the last three nights. Most of the time it says working and I don't really see a Speed. Sometimes 5Mbit. Actually the link is up to 10Mbit. We don't have the same speed on both side. Actually this is a test side.

Is this normal for this speed?

Thanks,

[fafa24]

An additional note, the first VM is complete, it says SUCCESS, now working on the 2VMs, which is on of the bigger ones in this chain.

[Shestakov]

Do you use WAN acceleration for those jobs? Note that for the first run accelerator makes cache population, next runs should be faster.
Was the run incremental or full? Thanks!

[fafa24]

Yes I run WAN accelerator on both side. The run was full without any other incremental.

[Shestakov]

The run was full without any other incremental.

Means seeding has not been done successfully.
As a result of seeding, VBR uses the backup file that you have created and moved to the target backup repository as a "seed", or starting point, for the backup chain produced by the initial backup copy job. And when a new restore point for the VM is available on the source backup repository, VBR transfers only incremental changes and store them next to the "seed".

[fafa24]

Hm, I created a normal backup job at the source site which produced a vbk. I move the whole folder with the vbm and vbk on a USD drive to the target repository offsite. Then I created a backup copy job using the repository on the target and mapped to the backup file which showed just fine. I populated then the global cache, selecting the target repository on the target WAN accelerator.

What went wrong?

Thanks,
Edy

[Shestakov]

Looks like you did everything as you should.
If you copied the .vbk to the target repository, rescan the repository, and map the .vbk on the target, next run of the backup copy job should create .vib with incremental changes only.
If you did everything correctly, but mapping doesn`t work, you can contact Veeam support team for the further issue investigation.
Thanks!

[fafa24]

I was a bit impatient. The seeding works, but it is slow. It took one week to finish the first pass. But this is a test side with a slow WAN link.

I have a new question. We will soon receive our Windows server for the production offsite. I intend to keep the server a few days first in the headquarter to populate it with the all backup copy jobs. Should I use the WAN accelerator inhouse as well, so the WAN cache will be build too? My understanding is when I move the server to the offsite location I would only need to modify the repository with the new address and backup copy jobs I would not need to modify since the repository is still the same. Is that correct?

Thanks,

[Shestakov]

Should I use the WAN accelerator inhouse as well, so the WAN cache will be build too?

It makes sense. You can either run a backup copy job with WAN acceleration or just populate cache.

VBR jobs use FQDN to recognize the server, so after moving the server offsite you should be fine.

[fafa24]

Hello,

I asked this question. Is it still with Version 3 only possible to seed a backup job and not a backup copy job? Anything changed how Version 9 works? We have a backup copy job pointing to our offsite repository which produces errors. I want to create a new backup copy job and seed the backup.

[Shestakov]

Hello,
It is possible to seed backup copy job in v9.
Thanks!