Hi, I've gone through the link provided by Luca. We're currently running on FC 8gb SAN & each of the Veeam backup is going through SAN mode where the speed is at satisfactory level. We'll be setting up a new nutanix cluster of 4 nodes using vsphere 5.5. Each nodes has 2 x 10Gb NIC to a pair of 10Gb switches. I'm a bit worried on the performance expectation using Direct NFS. Here's our setup.
1) Physical Backup server with Dualport 10Gb nic (each port going to a pair of 10gb switch to read NFS datastore)
2) Physical vCenter server (only has 1Gb ethernet production & qlogic 8gb Dualport Fibre Channel HBA going direct to an external FC Storage as repository)
3) 4 x Nutanix block node (each has Dualport 10Gb NIC going to a pair of 10gb switch)
4) A pair of 10gb switch (for redundancy)
5) Currently have around 50 VMs in total & backing up half of them daily/weekly/monthly
I've few questions from this setup.
1) Is this a supported architecture by Veeam?
2) Do I need to setup an internal vswitch on each ESXi hosts?
3) Does each node need an additional dedicated 10Gb NIC for backup traffic?
4) Does the restore uses the NFS mode as well? If not what is the requirement to use NFS? (NBD is usually slow)
5) And if it's not too much to ask, really appreciate if you could look into supporting AHV.
KVM is really gaining popularity with the openstack cloud.