Hyper Instance vm recovery and Migrate qns

[sacy]

Hi all

Im testing out Hyperv Instance vm recovery then migrate the instance vm to production. anyone notice when the migrate to production veeam using Production network? veeam traffic don't isolate within the configure backup network?

1. hyperv host are on 192.168.1.x/24 segment (production) / 192.168.2.x/24 (Backup network)
2. all veeam nodes are on 192.168.2.x/24 segment (Backup network)
3. on Veeam configure Preferred Network 192.168.2.0/24
4. allow the needed firewall ports for veeam's to 192.168.1.11

When running Instance vm recovery the mounting look running within 192.168.2.0/24 .
But when Migrate to Production trigger. veeam sent the data from 192.168.2.x to 192.168.1.x? as i can see on the firewall both production interface and backup network interface has the same amount of traffic loads. like backup network sent 300MB/s and Production receive 300MB/s.

[Andreas Neufert]

The preffered network setting is for communication between Veeam data movers. For example between the host at backup and the repository.

In your case the data flow is from mounted data to the host with Microsoft APIs. I am wondering why this data flow is not within the host itself. Do you migrate the server at same time to another host or different storage than the one selected at instant recovery?

[sacy]

not too sure what has happen. it tested on the same hyperv host. but if i do entire vm restore, the traffic are isolate within the backup network segment 192.168.2.x.

what i do.
1. run instance restore for JUMP-01 back to LAB-HYV-01
2. after JUMP-01-Restored are power up on LAB-HYP-01, then continue with the Migrate to production option under Instace Restore Job.

What i notice are when the Migrate to Production triggered, both my interfaces on the the virtual firewall has the same amount of traffic.
not sure is it netowrk configuration issue like dns?
https://youtu.be/_KFeMoEclnI

[sacy]

instance restore are mounted to the hyperv host via 192.168.2.x. but "migrate to production" step require to communicate over 192.168.1.x for hyperv host? possible due to i enroll the hyperv host via 192.168.1.x? but the hyperv host do have that 192.168.2.x interface.

guess i got to baseline expectation with my customer on instance recovery. or enroll the hyperv host to veeam server via backup network segment.

[johan.h]

Microsoft is introducing network intents as part of the Network ATC feature in Server 2025, but if for example data flow is happening from the Hyper-V host to itself, preferred networking is only honored by Veeam components.

What I think might be going on is that as migrate to production and IVMR are using a filter driver on the host side - we obviously don't want to mess with networking rules on the Hyper-V host .

https://learn.microsoft.com/en-us/azure ... ?tabs=22H2

[sacy]

thanks for the reply. ya. guess i better get my customer to use all backup network. and have a backup vlan on the hyperv host, and enroll the hyperv host via the backup vlan, lesser firewall rule to delt with. as im not able to explain why instant restoring of data need to sent over to the firewall.