Solved

Failed to deploy kubernetes cluster (Failed to deploy ntnx dvp)


Badge +2

Hi,

I’m trying to explore kubernetes by deploying Development Cluster. But it only stuck at 8% and eventually failed with notification (Internal Error : Failed to deploy ntnx dvp).

Here is my detail

  • Segment 192.168.2.x for Guest VM,PC,AOS and AHV.
  • Segment 192.168.3.x for Kubernetes cluster itself (DHCP Enabled)

Each segment can ping each other and all those segments are able to access the internet.

Here is the detail of the temporary deployed Karbon VM (After failed deploying, the VM got deleted automatically)

I’ve been reading someone posted same problem like mine, and it solved because of the VM previously unable to access the internet. But mine is able to access internet, PC and the Internal Nutanix.

Perhaps someone can enlighten me what did i do wrong?

-Able to pull newest docker image
-Able to ping Prism Central (192.168.2.11)
-Able to ping CVM (192.168.2.2)
-Able to access internet
Deploying Cluster stuck at 8%
Deleting VM after failure deployment
Notification Internal Error : Failed to deploy ntnx dvp

 

icon

Best answer by JoseNutanix 7 April 2021, 18:03

View original

This topic has been closed for comments

11 replies

Badge +2

Just adding some information, I just realized that when i want to deploy Karbon Cluster, i need to add the Data Services IP Address. How to enable it? Since i have no any external storage outside from my Cluster. I don't need external storage array to be added as a Data Services IP to my cluster?

Userlevel 4
Badge +3

Log in to the Prism Element on your cluster. Click on the cluster name located in the top left corner and set ISCSI Data Services IP if it’s not defined.

 

Badge +2

Yeah i know how to enable the Data Services. But what i want to ask is, what is this Data Services? Is it like an external storage array outside of my cluster? Like rubrick, Veeam, pure storage, etc? If that's the case, i wont be able to enable it since i dont have any. Or if its a VM that i need to create first inside the cluster, what is the detail of the vm?

Userlevel 4
Badge +3

It’s an ability to connect iscsi storage served by CVMs from your nutanix cluster to VMs, or containers. In case of k8s you will connect PVs using this data services. 

Badge +2

Thank you for answering my question. So eventually i can’t deploy Kubernetes cluster right now since i dont have any external iscsi storage array outside from my nutanix cluster?

Userlevel 4
Badge +3

No, your Nutanix cluster will act like iSCSI array. For this feature you need to set iSCSI Data Services IP. Read this prerequisites: Meet the indicated requirements prior to enabling and deploying Karbon

Badge +2

I’ve set the Data Services IP like shown below. I take unused IP on that segment to be as Data Services IP.

Data Services IP

And when i deployed the kubernetes cluster, it failed but the workers vm, etcd vm and master vm has been created like shown below

KBs VMs
And this is the iSCSI VG
This is the failure note when deploying KBs Cluster

 

Userlevel 4
Badge +5

Hi @KarangDika  Can you share what IP subnets are you configuring for the pod and service networks?

Badge +2

Hi @KarangDika  Can you share what IP subnets are you configuring for the pod and service networks?

Hi Jose,

 

I have 192.168.2.x segment for our current Guest VMs, CVMs and Hosts, PE and also PC

I have 192.168.3.x segment for my KBs cluster with DHCP Enabled from PRISM.

 

While the network pod i use default one from Prism Central that suggested me

Net Provider : Flannel

Pod : 172.20.0.0/16

Service : 172.19.0.0/16

 

I havent setup this network on my router and switch. Do i need to setup this network on my router first? Or i just change it into my current segment 2.x or 3.x?

 

Sorry if there is a silly mistake that i did.

 

Thank you.

Userlevel 4
Badge +5

Hi,

No, you don’t need to configure them in your router, they are internal to Kubernetes. The reason of asking for them was to check you weren’t using the VM or K8s segment as pod or service ranges. 

At this point I’d suggest you open a case with support, it’s going to be faster than trying to work out here what’s going on that is making your deployment to fail.

Regards,

Jose

Badge +2

Hi,

No, you don’t need to configure them in your router, they are internal to Kubernetes. The reason of asking for them was to check you weren’t using the VM or K8s segment as pod or service ranges. 

At this point I’d suggest you open a case with support, it’s going to be faster than trying to work out here what’s going on that is making your deployment to fail.

Regards,

Jose

Ahh i see. So from the whole settings, my KBs should be deployed normally right? But something wrong that prevent it.

 

Okay then, thank you @JoseNutanix i will open ticket then.