Edge Transport Node - Registration Timeout error

[Biowarez]

Hi,

 

I am doing a project at work with NSX-T and I went ahead and bought a server so I can run a lab at home. I am having issues with the edge transport nodes getting to a success configuration state, it hangs around on registration pending and then ends up at registration timeout.

Things I have tried:

• -Rebuild the node 10 times
• Confirm I can ping both edge vm nodes from my manager
• Confirmed 1ms latency between all 3 devices
• Increases the edge VM resources
• Tried to manually make the edge vm node join the manager on the CLI

Tonight I am planning to install an edge transport node via the OVF file rather than initiate it from the manager, not convinced it will make a difference, but I am unsure how to go about troubleshooting this.

Does anyone have any suggestions? Not sure why this is proving to be so problematic, this stage seems simple, provide network connectivity between the manager and the edge transport node and it just registers on the manager. Happy to provide any logs.

[p0wertje]

Hi,

Are the manager/edge-nodes/vcenter in the same subnet? (you don't have to, but just to get a better understanding of your setup.)
Is NSX already installed on the hosts where put your edge-node on ?(If so, check the DFW rules)

It sounds like a firewall or routing issue. i.e different subnets and your gateway applying nat.
Can you tell a bit more about your setup? A small diagram maybe?

[ShahabKhan]

Hi,

What is the error you are getting? I was building my home lab last week & faced similar problem. In my case, I was deploying the edge node on nested ESXi host. I also tried changing the CPU & RAM setting but without any success. After the troubleshooting, I found error was related to unsupported hardware. Somehow, the Edge node hardware was not supported when I was deploying it on nested ESXi. Then when I tried deploying it on my Server instead of nested ESXi, the deployment & registration was successful.

[Biowarez]

Hi,

 

Yes, my manager and edge transports are all in VLAN 10 / 10.0.10.0/24.

 

My server specs are:

 

LENOVO 11365K5

16 CPUs x Intel(R) Xeon(R) CPU E5-2650 0 @ 2.00GHz

64GB RAM

 

My host is running 3 clusters, one for NSX-Manager (single nested host), one for Compute (2 nested hosts) and one for NSX-Edge (single nested host), all running ESXI Ver 7.0

 

My Edge Transport Node UI page shows this:

 

 

 

 

 

My Alarms page shows this:

 

 

 

Ignore the transport node alarms, those hosts are not running for now.

 

I have uploaded the most recent logs from /var/log/vmware/nsx-syslog

 

Vcentre setup:

 

 

 

Attachment(s)

syslog.txt   578 KB 1 version

[ShahabKhan]

Can you please share the screenshot of Alerts from vCenter for edge nodes?

[Biowarez]

Hi,

I don't see any alarms on Vcentre for my edge VMs

 

I am firing up Vmware Workstation on my PC and will join to Vcentre and see if I can build the edge VM's. My home PC is a Intel 12th Gen CPU.

[JaSo2]

Hi,

I do not see a version you are installing - is it 3.2.x? Have you also tried 3.1.x? I'm wondering if it could be the huge pages problem again (in connection to your CPU) - check it here:

https://communities.vmware.com/t5/VMware-NSX-Discussions/NSX-T-3-2-on-vSphere-7u3c-Hugepage-issues-on-Edge-VMs/td-p/2894117

and possibly here:

https://communities.vmware.com/t5/VMware-NSX-Discussions/Nested-Lab-Can-t-deploy-NSX-Edge-Help/m-p/2907141#M14699

J.

[ShahabKhan]

What is the version of NSX-T? I had the issue with version 3.2.0 & 3.2.1 both.

[Biowarez]

Morning both,

 

I am using LE - NSX Manager/ NSX Global Manager / NSX Cloud Service Manager for VMware ESXi 2022-05-17 | 3.2.1 LE | 9.74 GB | ova

so it's version 3.2.1 LE.

Should I re-install my manager with NSXT ver 3.1 then?

 

 

[JaSo2]

well if you are not willing to tshoot the issue like described in the second post (my response), then you can try to install the 3.1 - it could be also a faster option in your case since I guess you do not have that much configured inside NSX-T yet anyway

J.

[Biowarez]

Hey,

I have rebuilt the manager and attempted to deploy a edge VM on Version 3.1 and I am getting the same symptoms. I attempted 2 builds and I watched the console for the edge build and one error message stands out:

failed to start edge datapath
see systemctl status nsx-edge-datapath.service

I tried googling this error, but there isnt much to go on, any ideas?

 

[ShahabKhan]

If possible, once try to deploy Edge node VMs on physical host instead of Nested ESXi host.

[Ivan Biasi]

Hi all
does anyone solved this error ?

I'm just facing the same issue using NSX 3.2.1 and deploying four new edges exactly in the same way I deployed others before.
Trying to register them also from the cli this error occurs:

nsxedge03> join management-plane 172.22.X.Y thumbprint 355920e4fc299ccbd66ed8d92fc01c70e9cae877a3eff729c95142390331e7d5 username admin

Password for API user:

Mon Feb 20 2023 UTC 15:27:49.359

% Unable to connect to the NSX API service