Veeam creates wrong static target when doing instant recover

[FrancWest]

Hi,

case number 02556472.

By design we have separated the iSCSI traffic for the ESXi hosts and the Veeam backup server in different vlans. The ESXi hosts have a connection to the storage array on two separate subnets for path redundancy. The Veeam server also has two separate subnets for its iSCSI connections to the storage array.

Now, when running instant VM recovery from a storage snapshot, it appears that Veeam creates a new static target on the ESXi hosts which is in the subnet of the Veeam server and which isn’t accessible by the ESXi hosts. This is obviously not correct.

Why does Veeam do this? The ESXi hosts are already configured with the correct static targets and can connect to the storage fine. So why is Veeam creating a new static target then? See below:

Static target 10.1.x.x is being created by Veeam automatically. This is the subnet we have assigned to the Veeam server to connect to the iSCSI storage, not the ESXi hosts.

But the ESXi hosts don’t have vmkernel adapters configured to access the 10.1.x.x subnet

So my assumption is that Veeam instructs the ESXi host to use the 10.1.x.x target to connect to the storage array, but since that subnet isn’t configured as a VMKernel adapter, it fails to connect and Instant VM Recovery fails.

Why is Veeam creating a wrong static target that isn't accessible by the ESXi hosts and why isn't Veeam using the existing configured static targets on the ESXi hosts to connect to the snapshot? These targets are working fine, since the snapshots are from production luns the ESXi hosts already are connected to.

Frank.