Backup copy job to 2nd location

[btmaus]

Hi all,

I currently have my Veeam backup jobs keeping 7 daily copies with Forever Forward Incremental on my repository in Site A.

I then have a backup copy job for GFS, keeping 4 weekly, 12 monthly and 7 yearly copies of the above data on the same repository in Site A.

I have a VBR server and two Proxies in Site A, and two Proxies in Site B.

I now want to also copy my GFS backup jobs (weekly/monthly/yearly) to a repository on Site B for data redundancy purposes.

How should I go about configuring this?

[Shestakov]

Hi Glenn,
Do you want to copy all existing files + new files or you just want to start copying new files to the site B?
You can copy the existing files manually and seed them.
Thanks!

[btmaus]

Hi thanks for the reply.

I am looking for advice on the best way to set this up. Copy just the new files would be fine, otherwise don't mind copying everything that is there plus new.

Do I need to create a 2nd backup copy job, that will make a copy of the 1st backup copy job? What would you advise?

[foggy]

The second backup copy job will create its own GFS restore points (remember that backup copy jobs do not simply copy files, but create restore points synthetically). Another approach is scheduling some script to simply sync GFS restore points (VBK files) to the second location.

[Shestakov]

Although, as foggy said second backup copy job will create it own GFS restore points, if you set same GFS schedule for both jobs, restore points will also be same.
Another option is to use file copy job to simply copy those backup files.

[rreed]

A "one-to-many" backup or copy job would be an awesome feature. Like the OP, we also need to be able to backup to location A, then once complete copy that same backup to both our dedupe devices and tape (whether location B or otherwise).

[Shestakov]

There is an option of secondary target for backup job, which can actually solve the OP`s and your issue. You can just point the source job to 2 backup copy jobs with different destinations.
Thanks!

[rreed]

Oooo winning!!! Thanks Shestakov! Question(s) though, do I recall seeing that Copy Jobs do not currently (v8) copy multiple VM's concurrently? And this is going to be fixed in V9? I've setup a test job as described; main Backup job goes to repository A, Copy (tape) job 1 goes to tape, Copy job 2 goes to dedupe repository (being able to order which happens first would be nice, or even parallel copy jobs FTW). Looking good but I notice the copy job is copying the first VM but the second one is pending, waiting for another task to complete. Admittedly, I do need to read up on exactly how backup copy jobs work and what to expect.

[Shestakov]

Hmm..
Actually both jobs should work simultaneously with no issues, because both of them just read the source backup file. I`m not sure there will be any related changes in v9.
Could you describe the behavior you observe, the source job finishes, 1 target job starts and another doesn`t? Or it starts as well but doesn`t copy the restore points?
I`ll try to test the scenario with QA team. Thanks!

[foggy]

In v9, you will be able to enable per-VM chains on backup repository, allowing multi-stream copy of data (hence, concurrent VM processing).

[rreed]

Sorry Shestakov, probably premature to report asynchronous copy jobs. I assumed it worked that way by looking at my Backup job properties and seeing my two copy jobs but no option to order. I haven't yet watched it run on its own to see what happens. User training, my fault!

And yes foggy, I am completely thrilled w/ real multi-streaming capability of v9. I'm in the process of shoring up our backup network infrastructure to make sure we're hopefully ready for it. Win!

[btmaus]

There is an option of secondary target for backup job, which can actually solve the OP`s and your issue. You can just point the source job to 2 backup copy jobs with different destinations.

Thanks I will give this a try. I think when you create a backup copy job, and specify the backup job as a source, it adds itself to the secondary target anyway?

Also, I don't have Enterprise Plus, so I can't use WAN accelerators. So I have some questions around throttling the network traffic that I hope you can assist me with:

Site A VBR and two Veeam proxyies are in 192.168.1.0 range. They each have a 2nd NIC to the 10.0.0.0 range for backup traffic to the NAS repository which is also in this same range. Site B has two proxies in the 192.168.2.0 range. They each have a 2nd NIC to the 10.0.1.0 range for backup traffic to the NAS repository which is also in this same range. The backup ranges (10.0.0.0 and 10.0.1.0) are not routable so the devices cannot communicate with each other between site A and site B, they have to communicate via the local Veeam Proxies at each site.

What is the correct way of doing this? I see there is an option on the VBR server, but also options on the Veeam Proxy servers as well.
If I specify the network throttling at the VBR server (Global setting), which ranges should be specified assuming I want to copy the backup jobs between both sites?
Which Veeam servers actually copy the traffic between the two sites? Is it between the proxy in site A to the proxy in site B?

Thanks!

[Shestakov]

Thanks I will give this a try. I think when you create a backup copy job, and specify the backup job as a source, it adds itself to the secondary target anyway?

Yes, when you choose backup job as a source of backupcopy, the copy job is being added as a secondary destination of the backup automatically.

During backup copying data is moving from source backup repository to the target backup repository over the network, proxies are not involved.
Throttling rules are to be applied between two Veeam agents running on the repository servers. You need to specify their IPs.
Thanks!

[btmaus]

Shestakov, my repositories are physical NAS appliances, they are not Veeam components. How is the data copied between them using Veeam?

[veremin]

Between gateway servers chosen in the setting of CIFS repositories. Thanks.

[btmaus]

Thanks, so it looks like it is the Veeam Proxy servers then, as they are the only servers i can choose if I manually specify the gateway server.

[foggy]

In this case traffic will indeed flow between those gateway servers.

[btmaus]

Ok so just doing a test on this. I had network traffic throttling setup in Veeam for the Proxy Servers in Site A to the Proxy Servers in Site B.

I started my job, and it kicked off at full speed. Looking into it further, the traffic is actually being sent from the VBR server in Site A, not the Veeam Proxies.

I edited the Network Traffic rules to include the VBR server, and then I see the "Network traffic throttling is enabled" message in the backup copy job, and the traffic slows down to what I specified.

So why is the traffic being sent from the VBR server?

[veremin]

It seems that no specific servers are specified as gateway servers, so those are selected in automatic fashion. By default, the role of gateway server is played by backup server, thus, the behaviour you observe. Thanks.

[Shestakov]

Glenn, as was said previously, Veeam proxies should not participate in backup copy job data transfer, unless you have gateways on the same machine with the proxies.
So the behavior you observe is expected.

[rreed]

By any chance, in v9, will the automatic gateway selection become smarter to select the nearest gateway server? Similar to the way proxy server selection is done?

[Shestakov]

As far as I know, there will be no changes regarding backup copy gateways and throttling. However we will take your opinion into consideration.

[rreed]

Excellent, many thanks. So what about the proxies in regards to backup copy? I see currently in v8 the B&R server handles the data moving in a Copy job - or is this what is discussed in regards to Gateway server for CIFS shares? If we have a fleet of proxies at the site we have to make the difficult decision which single one to choose, rather than it being able to use more than one like Backup jobs do?

[Shestakov]

You can either manually choose a gateway server for the repository or leave automatic selection and let VBR choose.

[rreed]

Thanks Shestakov. Is the automatic selction process smart enough to pick a gateway server for the given location? Say, for us, we have data center A and data center B. With source VM's/storage, proxies, and repositories all local to each data center, if I leave it to automatic will it know to select a gateway server for its correct location? Or would the random process potentially choose a gateway server for location A from location B, doing all the work across the WAN when it could have been doing it all locally instead?

[rreed]

And to help clarify for us (sorry, no intention to hijack thread), we perform all our backups locally to each data center. We do not backup across the WAN. In short, I'd like to feel confident B&R knows to keep all its traffic local. If it doesn't, I need to make sure I'm configuring correctly to remain w/in each location. If that makes sense.

[foggy]

The gateway server is picked at random (and per job session, so different gateways can be used for different job runs). So to keep traffic locally, you need to explicitly select the right gateway in repository settings.

[btmaus]

Is it possible to have a Server be a Veeam Gateway server for CIFS, and not be a Proxy?

[foggy]

Yes, can be any Windows server added to the Veeam B&R console.

[btmaus]

Just revisiting this one again.

I backup locally to Dell DR6000 appliances in Site A and Site B. The VBR server in site A cannot see the Dell DR6000 in Site B, it has to access it via one of the two Proxy/Gateway servers in Site B. I have two Proxy/Gateway servers in site A as well, that access their own local DR6000.

What do you think will be the best method of replicating my backup copy job data from Site A to Site B? Should I use the builtin Dell DR6000 replication, or the Veeam replication?

If I use Veeam, I need to specify a dedicated gateway server on both sites for the backup copy job data, so I can limit the network traffic. However, if I specify the gateway server, then Veeam will always use that server for the backup copy jobs itself, potentially overloading it?