Question

First install of CE fails - Network issue unable to route to 192.168.5.2

  • 8 September 2019
  • 4 replies
  • 203 views

Hi Everyone,

I've got 3 old Lenovo Tiny computers that sort of meet the bare minimum for running a little home-lab nutanix setup. I've been using them for months at home for my esxi setup but with a new AHV based rollout being planned for a few of our clients, I thought I'd get ahead by running up my home lab to get me familiar with Prism and migration from a few hypervisors (esxi and hyper-v) to AHV.

The lenovo tiny's are physically hardware configured like this:
  • 256GB M.2 SSD
  • 1TB 2.5" HDD
  • 64GB USB
  • 16GB RAM (though if proof of concept works, i'll go replace those with 16GB modules making 32GB per node)
I've built this as a bare metal. No nested virtualisation here.

Network assignment is a new virtual network, it doesn't really matter but i've got a cisco switch and a Barracuda CGF router. VLAN16 is my lan. Router is on 10.135.16.253.
I've got a Ubuntu machine on this network, it's able to route out to google browse sites, and even able to ping the host

My IP Assignment goes like this:
Host:
  • Host IP: 10.135.16.1
  • Host Subnet: 255.255.255.0 (24bit)
  • Host Gateway: 10.135.16.253
CVM:
  • CVM IP: 10.135.16.11
  • CVM Subnet: 255.255.255.0 (24bit)
  • Host Gateway: 10.135.16.253




The install gets to the point where it wants to autostart the NTNX-28028e87-A-CVM virtual machine. Then it eventually times out.



I looked at log file 'firstboot.out' and i can see the following lines:
  • INFO: Running cmd ['virsh autostart "NTNX-28028e87-A-VCM"]
  • INFO: Run ssh cmd on SVM
  • INFO: Running cmd ['/usr/bin/ssh -i /root/firstboot/ssh_keys/nutanix -o StrictHostKeyChecking=no -o NumberOfPasswordPrompt...
  • ssh: connect to host 192.168.5.2 port 22: No route to host]. Will retry in 5 seconds
  • INFO: Running cmd ['/usr/bin/ssh -i /root/firstboot/ssh_keys/nutanix -o StrictHostKeyChecking=no -o NumberOfPasswordPrompt...
  • ssh: connect to host 192.168.5.2 port 22: No route to host]. Will retry in 5 seconds
  • ...
  • ...



At first i had no idea this network was there and I wondered if i made a mistake, but it seems like no. That must be an internal network for CVM autopathing based on the googleing i did, i ended up here: https://vcdx133.com/2015/07/19/nutanix-xcp-deep-dive-part-17-cvm-autopathing-with-esxi/ and if i had to guess without knowing - i'd guess it's used for NOS to talk to it's VMs.

Anyway - clearly a network issue so I jump on to the machine with a keyboard and mouse (which i needed to install the machine anyway). First thing I did was try to ping 192.168.5.2. Destination host unreachable.



I check ip a and have a look at the network interfaces. Surprisingly theres 7 network interfaces, not including loopback. There's really only 2 that look interesting

br 0: IP: 10.135.16.1/24 brd 10.135.16.255
virbr0: IP 192.168.5.1/24 brd 192.168.5.255



I ping 192.168.5.1, pings back. Its myself, so i'm not surprised.
I ping 192.168.5.2, nothing


So i google again. I read that virsh is kvm's commands to check on guest vm's and in this case i want to see if the CVM guest is running so i go:

virsh list --all
code:
Id              Name                                      State
1 NTNX-28028e87-A-CVM Running



I dont' know anything about KVM but i use virsh -help and see there's a console command to connect to a VM's console so i make an educated guess

virsh console NTNX*
error: Failed to get domain 'NTNX-CVM.xml'
error: Domain not found: no domain with matching name NTNX-CVM.xml

Well.. not knowing kvm or nutanix yet, i can only guess that this is the configuration file used to setup the VM parameters or something since google tells me to edit that file. But i can run this command:
virsh dumpxml 1 | more
and that gives me an XML configuration for the CVM machine.

Now I've reinstalled a few times. So my Nutanix CVM machine name has been changing.

When i browse to /root i can see a file NTNX-CVM.xml and i can see there what looks to be a template.

But this is where I'm stuck.

I need some advice since i've got some surrounding information but no clear path forward. Even just advice for log files i haven't checked would be nice. I don't mind doing the legwork but this is my first time installing CE.

4 replies

Hate to shamelessly bump this, but all I think I need is a short list of logs that would be valuable to check.

So does anyone know where install log locations are?
CVM VM log locations?

Does this smell like the KVM machine hasn't booted? If so, and it says its running, how can i check?
Bump
Userlevel 1
Badge +1
Hey AidenClark,

You found the correct documentation about the 192.168.5.1 and 2 addresses. These are internal between the CVM (the .2) and the Host (the .1). The virtsh console command output is normal and that xml file is just that, the config file.

If you dump that XML file and look in the interfaces section, make sure you see an interface entry for br0 (your private network) and NTNX-Local-Network bridge=virbr0.

Try to SSH into the CVM Private IP you gave it ( 10.135.16.11) from the AHV host and see if it is actually running.
I had exact the same issue this week.
Difference is that I was installing NCE as nested AHV on ESXi. And AHV VMs were modified - added 3 more NICs to setup.
Reinstalled with only one NIC per VM, and issue did not appear again.
Maybe this will be a clue, but I can not be sure.

Reply