Veeam Replication from Backup and WAN Accelerator

[unsichtbarre]

Is there an advantage to Replication from Backup as opposed to Replication from Infrastructure VM when using WAN Accelerator, considering the Source WAN Accelerator does not process tasks simultaneously?

THX

[foggy]

I'd say it depends since these are a bit different scenarios. Some of the questions are if you do need to have a copy of your backups in the remote site, what kind of link do you have and what is the load you're going to assign to the setup in general (how many VMs, frequency of backup/replication, backup window, etc.).

[unsichtbarre]

Thanks foggy,

This is an IPsec VPN from Middle-East to USA. There are 57 VMs, and 22TB of used data. There are several "monster" VMs of 3-5TB each, the rest just a couple hundred GB each.

The IPsec pipe is 100Mbps on both ends, but I can only demonstrate about 30Mbps with iperf (related post: veeam-backup-replication-f2/long-distan ... 61809.html).

I am concerned primarily with diminishing results with the WAN Accelerator due to: a) only process one task at a time. b) need to run "disk fingerprint" per disk when using WAN Accelerator.

Any suggestions?

THX!

[foggy]

I'd test if the time needed for the WAN acceleration overhead will be less than it would take to transport the native workload over such a link. Keep in mind that you can use several source WAN accelerators to achieve parallel processing. Also keep in mind that with replication from backup you will also have backup copies in the remote site, which helps to comply with the 3-2-1 rule while using backup copy jobs to deliver backups offsite also allows using WAN acceleration.

[unsichtbarre]

In this case, the only Backup Repository is "Remote Side" (nearest the protected VMs, and the Replication from Backup sources the remote side Repository to create _replica VMs at the DR Site. Is this a mistake?

Here's the 10,0000 meter view:
1. PROD VMs to be protected run in Middle-East
2. IN PROD, Veeam Proxy/Repository runs as a physical server with (2) 8-core CPU, 512GB RAM and about 60TB HDD (think of it as a HPE Apollo)
3. IN PROD, the backup repository is on the Apollo
4. IN PROD, backup jobs run on the PROD Proxy site and target the PROD Apollo repository
5. DR Site is in USA
6. Link is a 100Mbps WAN (on both sides) IPsec VPN, transporting about 30Mbps max and unable to send ping larger than: ping -l 1410 -f from DR (USA) to PROD (Middle-East) or vice versa.
7. Veeam Server runs as a VM on the DR side - there is only a small (Default) Veeam Repository on the DR
8. Replication jobs are configured to source the PROD location Veeam Repository and Replicate to the DR Infrastructure as _replica VMs

[foggy]

Ok, so you're running replication from a remote backup, while I thought you first copy backups offsite and then use them as the source for replication. Both scenarios are possible, it's all up to your specific requirements.