Backup Copy job WAN target

[Kazz]

Hi,

We have 2 data centers connected over a slow WAN VPN link. (15Mbps/2MBps) Each site has a dedicated Veeam Proxy and a local NFS repo (Synology). Proxy at the HQ is a mount and a gateway server for the NFS datastore at HQ, proxy at DR is a mount and a gateway server for DR. VeeamBR at HQ manages backup jobs for both HQ and DR sites. All backup jobs are set to reverse incremental.
We also have one backup copy job that watches local NFS repository at HQ and copies VMs to DR as new backups are created. In the past we used Veeam to replicate VMs form HQ to DR, but recently decided to switch to a back up copy job to free up space on the VMWare datastore at DR. Backup Copy works as expected, but it seems to take longer to copy incremental changes to the DR site and we suspect it to be less efficient than replication. While setting up a replica job there was an option under storage to select if it was a LAN or WAN target, which I assume changed the block size of the backup files. We aren't seeing this option under back up copy job. Assuming it's using the setting from the source back up. If our assumption is correct, would it make sense to switch the storage optimization to WAN on the source backup job while it's targeting a local NFS repo? Would we decrease the performance of the local source backup and increase the performance of the back up copy ?

Thank you

[HannesK]

Hello,

which I assume changed the block size of the backup file

yes, that changes block size. With 256KB the incremental backups are around half of the default 1MB setting

We aren't seeing this option under back up copy job.

backup copy jobs inherit block size from the backup job. 1MB per default

would it make sense to switch the storage optimization to WAN on the source backup job while it's targeting a local NFS repo?

that would work, or you add WAN accelerators. Changing block size requires active full.

Best regards,
Hannes

[Kazz]

Thank you Hannes!

So sounds like we need to switch our source job settings to WAN optimized and run an Active Full.
Are we affecting our source backup job performance by switching Storage to be WAN optimized?
The copy job took 50+ hours to run initially, are we looking to have another 50 hour backup copy job after active full is ran?

Thank again

[HannesK]

Hello,
I would go with the WAN accelerator, but as you wish

Yes, it takes more RAM & CPU on the proxy / gateway server.

Yes, the backup copy job will take again the same time because it's the same amount of data. You can reduce the amount of data by increasing compression on the backup copy job (again, costs CPU)

Best regards,
Hannes

[Kazz]

We are on Ent Plus so WAN accelerator is available, but the overhead to manage 2 WAN accelerators at the current state does not make sense for us presently.

When it comes to WAN accelerators they require a dedicated server and physical box with ssd drives is preferred, but virtual would also work correct?

Our backup copy job contains 4 VMs (2 virtual appliances and 2 windows 2016 servers), the total backup size is 650GB and our change rate is aprox 20GB a day.

[HannesK]

yes, WAN accelerators also runs on VMs. But agree, for that amount of machines / data it's probably overkill.

20GByte over 15Mbit/s link should take around 3h. So that should work.