PVE can not create an NFS storage when using a second NIC

Hi to all

Anybody that can help me.

I see that PVE have a problem with the communication with the NFS storage when the PVE host have other IP address in other NIC for the NFS communication.

All my PVE servers have nfs-kernel-server installed and configured in a second physical and separate LAN.

For example:
Zone of Servers of Virtualization
PVE-1 on NIC1 = 10.1.1.1 -> LAN of PVE and VMs.
PVE-1 on NIC2 = 10.2.2.1 -> LAN of the Backups.
(Separate physical and logical networks for this host)

PVE-2 on NIC1 = 10.1.1.2 -> LAN of PVE and VMs.
PVE-2 on NIC2 = 10.2.2.2 -> LAN of the Backups.
(Separate physical and logical networks for this host)

Zone of NFS Servers
(without PVE Cluster)
PVE-NFS on NIC1 = 10.1.1.200 -> LAN of PVE.
PVE-NFS on NIC2 = 10.2.2.200 -> LAN of the Backups.
Also are separate physical and logical networks in this host.
(10.2.2.200 is the NIC where PVE-NFS must receive the data of backups)

When i try configure in the PVE GUI the NFS storage, i wait a long time, and nothing happens, when i try by CLI (from a Virtualization Server):
"showmount -e 10.2.2.20", after 30 seconds or 1 minute, I see the exportation.

Also, i have tried the command showmount between the distinct hosts PVE, and works well for the LAN communication but not for the LAN-OF-BACKUPS.

Comands as "ping" and "iperf" for the PVE-NFS Servers in the LAN-OF-BACKUPS give me good results.

Not matter if i have bond, bridge, or anything in the LAN-OF-BACKUPS, always "showmount" takes a while to show results.

This problem is present with PVE 2.3 , 3.1, and 3.3.

Can anybody tell me how fix it?