Possible or impossible?

[unsichtbarre]

Hypothetical scenario:

Is it possible to replicate a 50TB environment over an international 5,000 mile IPsec VPN (100mbps throughput 175ms. latency):

• 50 total VMs

• Source is a Backup *.VBK located at the remote side

• Destination is a *_replica

• Veeam Server is located at the replica site

• Multiple Proxies are located at the source site

• 2 VMs contain 15TB Disks each

• Source is vSphere/NetApp NAS

• Destination is vSphere/NetApp NAS

• Replicas were manually seeded by transporting a SAN to the destination site

Am I crazy to think this should be easy with Veeam? I am getting pushback from network managers saying the problem is Veeam and not the IPsec/network config!?!

THX for the input.

[EdgarRicharte]

I feel like it would be much faster to put that on physical media and ship that unit instead. I mean it should be doable. Its just going to take a while. It takes me days to evacuate 20TB disks locally.

[foggy]

Hi John, do I understand right that you've already done what is suggested above - moved the backup files physically to the remote site for replica seeding? In this case, you will only need to restore the data from backups and then continue replicating just changes, which should be possible - just test if the amount of changes fits the allowed backup window and check if WAN acceleration helps (available in replication jobs).

[unsichtbarre]

OK, sorry for ambiguity. I already seeded the environment using an old iSCSI SAN as a Veeam Repository. That part was no problem.

We have one VM that is causing trouble. It is taking 10 DAYS to calculate disk digests and so far, the process has failed at 90-100% on three occasions. The client claims it is a Veeam issue, we claim it must be a network issue.

We've found 2-3% packet loss across the VPN with wireshark
Ping -l 1472 -f <IP across VPN> returns a "request timed out" and not a "DF Bit" message (all VPN's I manage create a "DF Bit" message)
Ping -1 1410 -f <IP across VPN> works
There is about 180ms latency on the link

[foggy]

You've mentioned Veeam B&R server at the remote site - is it effectively used as a proxy server during replica job (digest calculation)?

[unsichtbarre]

I am not sure what we mean by "remote site" at this point.

B&R Server is at the"target" site with the *_replica VMs. B&R Proxies are at the "source" site with the production VMs and Veeam Repository which is also the source for the *_replica VMs.

We have specified the Target proxy in the Replication from Backup job, but there is no specification for Source proxy as it is Replication from Backup. We have the replica metadata stored on a repository at the Source site as is outlined in: https://helpcenter.veeam.com/docs/backu ... l?ver=95u4

[foggy]

And the target proxy is located in the target site, right? Otherwise digests calculation would be ineffectively slow.

[unsichtbarre]

Yes. Target Proxies are at the Target site, closest to the replicas. Replication metadata is stored on a Repository at the Source Side closest to the Production VMs and Source Backup Repository.

[foggy]

Then everything seems to be configured properly. For further investigation, I recommend contacting technical support.