Solved

Nutanix CE installs but CVM does not create cluster in vmware nested environment.

  • 8 April 2024
  • 9 replies
  • 129 views

Badge

Hi,

I successfully deployed three nodes as VMs in my VMware nested environment without any issues. All are pinging each other. CVMs are also pingable from each other.

But when I create cluster from a CVM, cluster creation failed, medusa and its below listed services are having DOWN status. 

Please help.

Regards,

Gurdeep Sandhu

 

icon

Best answer by JeroenTielen 8 April 2024, 14:56

View original

9 replies

Userlevel 6
Badge +8

What are the specs of the virtual machines and what is the spec for the VMWare server? 

And do you have a screenshot of error? 

Userlevel 4
Badge +4

Hello!

Please check my article on how to deploy nested CE 2.0: https://en.vmik.net/2023/09/ntnx-ce20-vsphere8-deploy/

 

In short, for each CE VM you need to add advanced parameter disk.EnableUUID = TRUE. Otherwise it can affect Medusa service during cluster creation process.

 

 
Badge

Hi,

I created 3 VMs on my ESXi Hypervisor with below configurations:

  1. 4 vCPUs (HW virtualization enabled)
  2. 32 GB memory
  3. SCSI Controller VMware Paravirtual
  4. HDD1- 16GB, thin provisioned, SATA controller 0, used for Hypervisor Boot
  5. HDD2- 200GB, thin provisioned, SATA controller 0, used for Data
  6. HDD3-500GB, thin provisioned, SATA controller 0, used for CVM
  7. disk.EnableUUID to TRUE
  8. 1 VMXNET3 Network Adapter
  9. on vSphere Standard Switch: Promiscous mode=Accept, MAC Address change=Accept, Forged Transmits=Accept

 

Userlevel 6
Badge +8

Your specs aren't enough (Boot disk is to small). Please follow my tutorial exactly and it will work: https://www.jeroentielen.nl/installing-nutanix-community-edition-ce-on-vmware-esxi-vsphere/

 

 

Badge

Hi Jeroen,

Big thanks. With you step-by-step guide, I am able to deploy the cluster.

I was able to login 1st time in Prism Element. Registered for Nutanix NEXT also. Now, I am facing Prism element login issue. Getting Error: Server not reachable. All CVMs are pingable from each other. 1st time login was perfect. 

Please help.

Userlevel 6
Badge +8

If you just started the cluster then wait a couple of minutes more. If the cluster was running for days try to give a cluster start command in cvm. 

Badge

Hi,

I deployed cluster a few days back. While login, throwing an error: server not reachable.

 

 

Why Cluster start process taking too long, It’s taking around 30 minutes or more to start all the services. Is this default behaviour or do I have to make some changes. Please guide.

Thanks.

Userlevel 6
Badge +8

That depends on the physical hardware. Can take a while to start things up yes. 

Userlevel 4
Badge +4

I agree with Jeroen. It can depends on the underlying hardware, especially on the disks.

Faced the same on non-nested CE. The simpiest indicator that something goes slow:

Check metadata ring - nodetool -h0 ring - nodes may be in the down state.

Check disks utilization and await time from CVMs: sar -dp or iostat

Reply