NSX-Hyperbus VIB is a key component of Kubernetes to NSX-T integration.
Yet the vib nsx-hyperbus does not seem to be published anywhere in vmware site. Looked in the LCP files for ESXi NSX-T and could not find it.
It looks like it isn't installed by default by the NSX-Manager when doing host prep.
With a standard installation of NSX-T, none of the hosts I have possess that vib installed and I cannot get it from anywhere.
When I try to deploy a new container app in kubernetes I'm stuck at ContainerCreation pending as my node-agents are constantly failing their liveness probe because the get hyperbus status is unhealthy.
When I go to the ESX server I just see an empty answer when I look for get hyperbus connection info, because the vib is nowhere to be found.
Even on the manifest file of the repositories in the NSX-M nsx-hyperbus does not exist.
Yet in vmware hands-on labs the vib is installed and exists. I just cannot figure out where to get it for my install.
Can anybody point me on how can I get that vib pushed over to the ESX servers where my k8s nodes are running?
The hyperbus is part of netcpad (/etc/init.d/netcpad) and should be pushed when you prepare your hosts as transport nodes. You can use nsxcli on the ESXi host to check its status (get hyperbus connection info). If you're saying that it isn't available, then you have a problem somewhere and I'd start with the logs on Manager and the ESXi hosts to see what happened there. Ensure that you're using supported versions of the components all the way around as well.
Thank you for your answer. Are you saying that the vib isn't part of the local control plane vibs that are packaged with nsx kernel files?
What I find strange is that if I compare my install with nsx-t hands on labs install I see:
- the esxi hosts in HoL do have nsx-hyperbus vib installed
- the manifest file in the nsx manager repository (in http://<nsx-manager-ip-address>:8080/repository/<target-nsx-t-version>/metadata/manifest) does show nsx-hyperbus vib in Hands on labs
- my install has none of these
Is there any specific know I should've turned on to say that my install should be "container-ready"? Licensing editions do not appear to indicate that.
Just an update.
Apparently the VIB nsx-hyperbus has disappeared from the lcp files since NSX-T 2.3 (was there in NSX-T 2.2 lcp file), but there's absolutely no notice on release notes or documentation on how did the behaviour change.
My get nsx-hyperbus connection is empty and due to this my node-agent containers are failing their liveness probes, which stops container deployment.
If anybody has an idea, it is more than welcome.
Did you ever figure out your hyperbus problems? I'm troubleshooting a similar issue.
My ESXi servers don't show anything when I run the "nsxcli -c get hyperbus connection info"
And my nsx-node-agent pods report "Unhealthy" when I run the nsxcli "get node-agent-hyperbus status"
This seems to have broken after we re-deployed our kube cluster, but was working before.