VMware Networking Community
orddie
Enthusiast
Enthusiast
Jump to solution

new nsx-t install showing DEGRADED

this is for the NSX Application Platform.

As far as I can tell, its the ANALYTICS services thats keeping the cluster from going green and adding more services. Not sure where to start trouble shooting on this one.

 

orddie_0-1650728733147.png

 

Reply
0 Kudos
1 Solution

Accepted Solutions
p0wertje
Hot Shot
Hot Shot
Jump to solution

You should be able to add

volumes:
        - name: containerd
          mountPath: /var/lib/containerd
          capacity:
            storage: 64Gi

under 'workers:'

As described in https://docs.vmware.com/en/VMware-vSphere/7.0/vmware-vsphere-with-tanzu/GUID-4E68C7F2-C948-489A-A909... 

Cheers,
p0wertje | VCIX6-NV | JNCIS-ENT | vExpert
Please kudo helpful posts and mark the thread as solved if solved

View solution in original post

Reply
0 Kudos
9 Replies
p0wertje
Hot Shot
Hot Shot
Jump to solution

Are all the pods running? (kubectl get pods -n nsxi-platform)
And check the statefulset.

kubectl get statefulset -n nsxi-platform
NAME READY AGE
druid-config-historical 1/1 43d
druid-historical 2/2 43d
druid-middle-manager 3/3 43d
fluentd 1/1 43d
kafka 3/3 43d
llanta-detectors 1/1 43d
metrics-postgresql-ha-postgresql 3/3 43d
minio 4/4 43d
postgresql-ha-postgresql 1/1 43d
redis-master 1/1 43d
redis-slave 2/2 43d
zookeeper 3/3 43d

Cheers,
p0wertje | VCIX6-NV | JNCIS-ENT | vExpert
Please kudo helpful posts and mark the thread as solved if solved
orddie
Enthusiast
Enthusiast
Jump to solution

looks like i have some pods that are not running.  how do I fix?

 

 

orddie_0-1650734117755.pngorddie_1-1650734121699.png

 

 

Reply
0 Kudos
p0wertje
Hot Shot
Hot Shot
Jump to solution

It kinda depends on why it crashes,
My best guess is that you did not assign enough storage to "/var/lob/containerd". You should use something like 64gb
See https://docs.vmware.com/en/VMware-NSX-T-Data-Center/3.2/nsx-application-platform/GUID-85CD2728-8081-... for the requirements

 

Where did you build napp on ? Tanzu? or just normal kubernetes

 

Cheers,
p0wertje | VCIX6-NV | JNCIS-ENT | vExpert
Please kudo helpful posts and mark the thread as solved if solved
orddie
Enthusiast
Enthusiast
Jump to solution

I used Tanzu.

You may be right! looking at usage, looks like each worker node got 15GB of space. I'm not sure why it set to 15GB of space. No policy is limiting this as far as I can tell.

orddie_0-1650741002598.png

this was the YAML I used to setup the cluster.

apiVersion: run.tanzu.vmware.com/v1alpha1
kind: TanzuKubernetesCluster
metadata:
name: nsxappsclust
namespace: nsxtapps
spec:
distribution:
version: v1.20.7
topology:
controlPlane:
class: best-effort-small
count: 1
storageClass: tanzubasic
workers:
class: nsxenv
count: 3
storageClass: tanzubasic
settings:
network:
cni:
name: calico
pods:
cidrBlocks: ["192.168.120.0/24"]
storage:
defaultClass: tanzubasic

 

and this is the template showing resources.

orddie_1-1650741143779.png

 

Reply
0 Kudos
p0wertje
Hot Shot
Hot Shot
Jump to solution

You should be able to add

volumes:
        - name: containerd
          mountPath: /var/lib/containerd
          capacity:
            storage: 64Gi

under 'workers:'

As described in https://docs.vmware.com/en/VMware-vSphere/7.0/vmware-vsphere-with-tanzu/GUID-4E68C7F2-C948-489A-A909... 

Cheers,
p0wertje | VCIX6-NV | JNCIS-ENT | vExpert
Please kudo helpful posts and mark the thread as solved if solved
Reply
0 Kudos
orddie
Enthusiast
Enthusiast
Jump to solution

Looking through the links you gave me, my setup is not right at all.  looks like I hurt myself with my YAML file.  Working on another version now and will redeploy.

 

Thx MATE!

Reply
0 Kudos
orddie
Enthusiast
Enthusiast
Jump to solution

I'm actually seeing I need to add it for the control as well.  do you agree?

orddie_1-1650742101502.png

 

 

Reply
0 Kudos
p0wertje
Hot Shot
Hot Shot
Jump to solution

Yes. according to the documentation, you should add that for etcd on the control.
I had some issues doing it. And napp was running fine with only on the workers

 

Cheers,
p0wertje | VCIX6-NV | JNCIS-ENT | vExpert
Please kudo helpful posts and mark the thread as solved if solved
Reply
0 Kudos
orddie
Enthusiast
Enthusiast
Jump to solution

adding a claim of 65Gi was correct.

 

this was the first time it ever deployed without me having to hit retry during install.

will be working on service activation next.  Thanks for getting me this far!

Reply
0 Kudos