NSX-T / 2 pnic / Edge Deployment

[alsmk2]

I'm hoping someone can provide some help with this issue. I've deployed an NSX-T manager cluster on to hosts that have only 2 physical nics. I've migrated all vmkernels over to switches/ segments on the single N-VDS that has been deployed without too many issues.

I'm now trying to deploy an Edge VM on top of the N-VDS following the steps in this video:

Demo: NSX-T Data Center Edge VM deployment on N-VDS Logical switches - YouTube

I can't get this to deploy cleanly at all. If I try leveraging the deployment via postman, I get the following error:

"[Fabric] Network interface '6f675985-3259-4ed2-b7e2-57ddf5d6ec29' is either an overlay logical switch or in an unaccessible Transport zone. Only accessible VLAN logical switches are supported."

If I deploy manually via the edge ova and register the node against the management plane, it goes through, but on the nodes page I have the following under configuration status:

"Edge VM hardware or Underlying Host configuration is not supported"

In terms of the setup I have the following:

Transport Zones:

1. tz-overlay (n-vds1)

2. tz-vlan (n-vds1)

3. tz-edge-uplink1 (n-vds2) <-- Has a status of unknown

4. tz-edge-uplink2 (n-vds3)<-- Has a status of unknown

Segments:

1. Management-LS (tz-vlan) (ESXi TEP IP also on this vlan)

2. Edge-Transport-LS (tz-vlan) (EDGE TEP here - defined different vlan)

3. Edge-Uplink01-LS (tz-vlan)

4. Edge-Uplink02-LS (tz-vlan)

5. Edge-W-Uplink01-LS (tz-edge-uplink1)

6. Edge-W-Uplink02-LS (tz-edge-uplink2)

For the OVA I select the following:

1. nic 0 - Management-LS

2. nic1 - Edge-Transport-LS

3. nic2 - Edge-Uplink01-LS

4. nic3 - Edge-Uplink02-LS

As mentioned earlier, via OVA the node deploys and I can join it to the management plane, but I get the message highlighted above. If ignore this and try to configure it as a transport node, after I've configured the various n-vds settings, I get an additional error about the edge pnics not being prepared (I've deleted the edge node, so can't remember the exact error).

The error I get from the API call doesn't really help - the referenced network is already in use by the ESXi hosts and looks absolutely fine. Can anyone please advise on what could be the issue?

[alsmk2]

Further update, I can see the following error when I ssh to the Edge VM:

NSX Edge configuration has failed. The host does not support required cpu features: ['aes']

According to documentation, the solution here is to restart the datapth service (dataplane?) and reboot the appliance.

If I stop the datapath service, it won't start again. Still stuck, but edging closer.

[daphnissov · Accepted solution]

That error from the edges means that the AES CPU instruction set is not available, which is required for edge functionality. You probably have EVC set to a level lower than that which is required (Westmere).

---------------------------------------------------------------------------------------------------------

Was it helpful? Let us know by completing this short survey here.

[alsmk2]

You were 100% spot on - EVC enabled to facilitate migration from old cluster.

Removed it and the edge deployed without any issues. Thanks for posting