v6 - How to have a local and off-site backup copy?

[lsittech]

Thanks jpeake and Vitaliy. Secondary job sounds easiest and that way its all controlled in the one interface.
Sorry for late reply, been off working on other issues.

[samwren]

Hi,

We are using Veaam Backup and Replication 6.5.0.109 (64 BIT) and this is running daily backups of all our Servers. We realized we need to take these off site in case of an office fire etc. So we have been moving the backup files onto a 2TB external USB device. However, the Full backup vbk file is over 884GBs in size and this is growing by around 2GBs a day. When I manually copy this and the evening's Reversed Incremental backup VRB file this can take up to 8 hours to be performed.

Does anyone recommend a better way of getting this data onto the USB device? Obviously at present this device is taken home each night (we are soon to use a Fire safe) so we cannot use this as the backup to directory when Veaam runs it's jobs. The copy time is becoming an issue.

Unfortunately trying to upload to the Cloud is not feasible as the size of the backup files are over 1.2TBs and growing. I am using Extreme Compression but this does not seem to be making much of a difference.

Thanks in advance.

Sam

[Yuki]

I see that many suggest using R-Sync, but how do people actually set it up with reverse incrementals??

generally RSYNC will not detect renamed files (which is what i believe happens with reverse incremental where changes from last backed up are added to last backup file (the one that is always a full backup) and the difference is moved out into an incremental file for that day (previous day).

Also initializing RSYNC over large repository with TBs of data (our single full backup file is 2.4TB after compression)+ incrementals (100-200GB) takes very long time to hash and check on both ends.

I've tried always sync which is supposed to track renamed files, but it also takes very long time to start sync.

It seems that rolling a second backup job to remote site is the best way to go, but then we are hitting production servers twice!

Solutions?

[Vitaliy S.]

Hi Yuki,

You can use this registry key to enable old-style reversed incremental VBK file naming convention. In addition to this please take a look at these threads for additional info: V5 VBK file names and Offsite backup script.

Hope this helps!

[dschrader@iu17.org]

Hi Guys,

I currently have to separate sites. One site is my main production site and my other site has one server that has a DC and an application program on it. I have my Veeam server in my production site running on a Dell R710 with 2TB drives Giving me 12TB before Raid with a 20GB link directly to our core switch. In my other site I have a Dell R610 with a MD1000 with 2TB drives giving me 30TB before RAID. Here I have a 200MB link to my production site.

My idea was to use my production Veeam server as a "Short Term Backup" location and the R610 as my "Long Term Backup" location. Veeam is only installed in my production environment on my R710.

Is there a way using one backup job to put the backups in two separate places? Also is there a way to tell on disk to store backups for 7 days and the other disk to store backups for 30 days for example?

During the day I wish to use my production Veeam server to do recoveries from (since it is faster and on site). I would rather not have two separate jobs processing the same virtual machine during the night. It just doesn't make sense to process the same data twice.

Any ideas?
Thanks,
Dan

[dellock6]

Hi Daniel,
at the moment Veeam can only save to one location per job. You can either run the same job twice saving to two distinct location, or use a third party replication software to copy backup files from the primary location to the second one. Your need is a common request for enhancement coming out in these forums, hope Veeam will listen

Luca.

[evander]

Hi - So I am backing up locally and then also replicating all my VMs to my remote DR site.
I am suprised this is not an option for more people on this thread. Am I missing something?

I realise you dont get the same compression during a replication but then again you want your DR site to be able to run your full environment should the Primary site be taken out completely.Right?

[foggy]

You're right, but take into account short retention period available for replication. With replication, you do have full environment ready for immediate failover, however backups at both local and remote sites allow for a much longer (and less regular) retention in a space-efficient manner. Although it's all about your backup strategy but, as mentioned above, backup is not a backup until you have at least 3 copies of it.

[rubeng]

Hi guys,

I have some questions regarding the situation where you need a local backup and an off-site backup location. There has been some interesting development since the last post I made about off-site backup location. By development, I am referring to the Backup Copy Job and the 'ForceCreateMissingVBK' registry key.

Our previous setup was like this:

Datacenter 1, Building 1 contains several virtual machines and a Veeam B&R v7 R2 server. Datacenter 2, Building 2 contains some non-important servers and a Veeam B&R v7R2. These two buildings are connected by 1 gig network. (Soon fiber). Veeam B&R in datacenter 2 was automatically copying every single backup file from datacenter 1, every day. We were never able to make the software we use for synchronization look for the changes in the .VBK files, so it copied the whole file every day, even if there were just small changes to it. This made the copy take ages, and we are nearly at a point where we cannot copy any more data cause of the time it takes. However, thanks to Backup Copy Job, this is not a problem anymore.

The Veeam B&R server at datacenter 2, containing a copy of all the data in datacenter1, also copies all the data it has synchronized onto rotating USB hard drives. This is of course also a time consuming operation. We were hoping we could make Veeam do this job as well. However, as long as the server in datacenter 2 is not the server containing the backup job, or the backup copy job, it’s not possible to synchronize the data onto the disk. Correct? I find this a bit strange, considering the Backup Copy Job in datacenter 1 is copying to a repository on the server in datacenter 2.
We could continue to use the solution we have used for years, raw file copying. But it would be nice if Veeam handled that job as well.

Here's an image of the setup, in case the explanation of our setup is a bit unclear:



Does anyone have a good solution to this, or know if it's possible to accomplish what we're trying to accomplish with Veeam?

I would like to say thanks in advance for all replies and tips


Best Regards,

Ruben Gilja

[veremin]

Hi, Ruben, you can try to

1) Create a backup job at the second site.
2) Map it to the files produced by backup copy job.
3) Select the backup job as a source for a backup copy job that points to rotated media.
4) Implement the regkey mentioned in this thread and see whether it helps or not.

Thanks.

[rubeng]

Thanks for quick reply v.Eremin

I will give it a try!

[rubeng]

v.Eremin, regarding the Backup Copy Job over LAN. Do you think we would benefit from Jumbo Frames? The setup I described over, will move a few TB every day.

[veremin]

The actual tests with "fake" jobs should give you the final answer. However, be aware to configure Jumbo frames properly - the statement from the recent forum digest (though, describing some other matter) is still valid here:

Remember that you should enable jumbo frames *everywhere* within the given network segment. Enabling jumbo frames can do you more bad than good if you are not diligent enough implementing the change.

Thanks.

[rubeng]

Thanks for quick reply!

We'll do some testing

[veremin]

Kindly, keep us updated about the results you get. Thanks.

[edfops]

I would love to here if adding this as a feature is still on a roadmap. Having to manage/use another tool just to fit a 3-2-1 backup strategy is quite a pain, and it's something a lot of people are trying to accomplish.

[foggy]

Have you heard about backup copy jobs available since v7?

[edfops]

Have you heard about backup copy jobs available since v7?

Foggy, if I take a primary backup to fast disk and then backup copy it to slower disk, I cannot then create a 2nd backup copy job pointing to the backups stored on slower storages.

My options are
- Perform a regular backup to the slower storage, which will have a huge impact on my production VMs during backups. I can then backup copy this to the 3rd location.
-Use a 3rd party tool to replicate my backup copies to the 3rd location.

It would be great if Veeam had a way to replicate files (and just their changes nightly) to a 3rd location in a backup chain.

[foggy]

Ah, I now understand your concern, didn't get it initially, my bad. Starting v8, backup copy job will be able to use backups produced by another backup copy job as a source.

[edfops]

Ah, I now understand your concern, didn't get it initially, my bad. Starting v8, backup copy job will be able to use backups produced by another backup copy job as a source.

This is HUGE. Very much looking forward to v8 alongside some of the other stuff coming with it.

Good time to ask, is there a full feature set listed for all the fixes/features/bugs resolved in v8 since it is a few weeks from release?

[foggy]

Will be published soon.

[edfops]

Will be published soon.

Testing in version 8 and it does not appear to be a direct option. I see the ability to pick a "source" repository, and this does in fact work, but requires I pick specific VMs from infrastructure.

If you select a specific source, but point your selected VMs at a backup job (so it can populate its VM list, you cannot choose a backup copy job for this), it will copy from the backup job repository and not the selected backup copy repository.

[Gostev]

Testing in version 8 and it does not appear to be a direct option. I see the ability to pick a "source" repository, and this does in fact work, but requires I pick specific VMs from infrastructure.

Yes, this is explained in the What's New document.

[MutazN]

Hello Guys,

I have the same concern here about saving a local copy and a remote copy by the replication job or any other way in V8.

please help me

Any help would be greatly appreciated.

Regards,

[Vitaliy S.]

Can you please clarify your concern? If you want to have two backup copies, then use backup jobs to backup your VMs locally, and then create backup copy jobs to send these VM backups to the offsite location.

[MutazN]

Thank you for your reply.

I need to create replication job from local host(ESXi) to the remote host (ESXi) and I need to save the same replication job to a local location.

Regards,

[dellock6]

In V8 you can now use a backup as a source for replication.
So one solution for you can be:
- backup saved locally for the primary location
- replication to remote site using backup as the source
In this way you only connect to production storage once, thus lowering the I/O you create there.

[foggy]

But if you need both local and remote replicas, there's no way to avoid creating two replication jobs (though both can use backups as a source).