Partial failover between vCenters

[bdelacour]

Hello,

I'd like to use Veeam to migrate my VMs from one vCenter to another on the same site.
I installed a Windows Virtual Machine with Veeam Backup & Replication and I am able to list my VMs.

As I have some virtual machines, I would like to migrate service per service using partial failover. For this, I need to setup Network Extension Applicance.
The problem is I apparently need Veeam Cloud Connect that I can't get without being a Service Provider.

Here is how my network works for information :
I use 3 VLANs and a firewall/gateway in a VM to manage communications. My services can, for example, have 14 VMs on VLAN1, 17 VMs on VLAN2 and 6 VMs on VLAN3.

Do you have a solution ?

Thank you !

[PetrM]

Hi Benjamin,

If I correctly understood the scenario you don't need to leverage Partial Failover and Cloud Connect but you should use Quick Migration job instead.

Thanks!

[bdelacour]

Hi Petr !

Thanks a lot for your answer, I didn't know about "Quick Migration", it can be a good solution !
I'll do some tests to measure unavailability time on our services using this method.

For my own culture, would it have been possible to migrate service per service by establishing a network connection between both vCenter without being a Service Provider as I first thought ?

Kind regards

[veremin]

Correct, as long as there is a connectivity in place you will be able to migrate VMs between vCenters, using either Quick Migration or replication coupled with planned failover. Thanks!

[bdelacour]

I think my question was not well written, sorry

Would it be possible to setup Partial Failover without Cloud Connect ?
That, if half of a VLAN is migrated, all machines can still communicate with each other even if they are not in the same vCenter.

Thanks !

[veremin]

No, there are no similar means available in VB&R. Thanks!

[bdelacour]

Hi again

I am testing the Quick Migration part and I've got two questions :

1- As we do in failover, is it possible to have a quick rollback in case of failure ?
I unchecked the option which deletes the VM after migration but it is renamed and I don't want to resume 200VMs by hand.

2- I was able to make a migration but, even if my DVS have the same name on both vCenter, VB&R is not able to configure network interface. It just appears disconnected with a blank adapter selected.
In Event, I am just able to see this logs :

Code: Select all

 Modified: config.hardware.device(4000).backing.port.portgroupKey: "dvportgroup-77" -> <unset>; config.hardware.device(4000).backing.port.portKey: "273" -> <unset>; config.hardware.device(4000).backing.port.connectionCookie: 1942004904 -> <unset>; config.extraConfig("tools.guest.desktop.autolock").value: "false" -> "FALSE"; Added: Deleted: 

And if I configure it by myself, it logs :

Code: Select all

Modified: config.hardware.device(4000).deviceInfo.summary: "DVSwitch: 41 a5 25 50 63 b7 73 95-f2 58 23 68 3a ad 7c 1e" -> "DVSwitch: 50 0d ee d2 11 93 59 30-10 4a 05 f9 12 2e 7f 06"; config.hardware.device(4000).backing.port.switchUuid: "41 a5 25 50 63 b7 73 95-f2 58 23 68 3a ad 7c 1e" -> "50 0d ee d2 11 93 59 30-10 4a 05 f9 12 2e 7f 06"; config.hardware.device(4000).backing.port.portgroupKey: <unset> -> "dvportgroup-45"; config.hardware.device(4000).backing.port.portKey: <unset> -> "256"; config.hardware.device(4000).backing.port.connectionCookie: <unset> -> 1059394735; config.hardware.device(4000).connectable.connected: false -> true; config.hardware.device(4000).connectable.status: "unrecoverableError" -> "ok"; Added: Deleted: 

I might create a new post for this problem if needed

Thanks a lot !

[PetrM]

Hi Benjamin,

1. There is no automatic rollback, Quick Migration allows you to migrate your workloads from one location to another at once. You may also take a look at replication and planned failover functionality (as Veremin has already suggested) if you want to be able to perform Undo Failover or Failback operations after failing over to a replica.

2. Since it's related to troubleshooting of a technical issue, I would recommend to raise a support request because we cannot fix technical issues over forum posts. Please don't forget to provide us with a support case ID.

Thanks!

[bdelacour]

Hi Petr,

Since I will use replication and planned failover because I must be able to failback, I will use the network remapping feature of replication jobs.
I have the same interface name on both side but at least, internally it will be able map them

Thank you to both of you for your answers !