No available interfaces for cluster creation

[ltcptgeneral]

From the documentation, it seems to suggest that using a separate network interface for the cluster is optional, but when I try to create a cluster, there are no available interfaces to use. The current deployment only has one interface which is slaved by the linux bridge. I can add another NIC in the future, but I would prefer to test the feature without having to add a bunch of NICs to each node. Is there anyway to use the existing interface to test this feature?

What I am expecting (from the docs):


What I am getting:

 

[fiona]

Hi,
what is the output of pvesh get nodes/localhost/network --output-format json-pretty and cat /etc/network/interfaces? Feel free to censor any public IPs, but in such a way that it's still possible to tell if two are the same.

 

[ltcptgeneral]

Hi,
what is the output of pvesh get nodes/localhost/network --output-format json-pretty and cat /etc/network/interfaces? Feel free to censor any public IPs, but in such a way that it's still possible to tell if two are the same.

[
   {
      "active" : 1,
      "autostart" : 1,
      "bridge_fd" : "0",
      "bridge_ports" : "eno1",
      "bridge_stp" : "off",
      "bridge_vids" : "2-4094",
      "bridge_vlan_aware" : 1,
      "families" : [
         "inet"
      ],
      "iface" : "vmbr0",
      "method" : "manual",
      "method6" : "manual",
      "priority" : 4,
      "type" : "bridge"
   },
   {
      "active" : 1,
      "exists" : 1,
      "families" : [
         "inet"
      ],
      "iface" : "eno1",
      "method" : "manual",
      "method6" : "manual",
      "priority" : 3,
      "type" : "eth"
   }
]

auto lo
iface lo inet loopback

iface eno1 inet manual

auto vmbr0
iface vmbr0 inet manual
        bridge-ports eno1
        bridge-stp off
        bridge-fd 0
        bridge-vlan-aware yes
        bridge-vids 2-4094

 

[fiona]

Well, there's no IP address assigned to the bridge. It should match the one configured in /etc/hosts.

 

[ltcptgeneral]

That seems to have solved the issue. However, I prefer that nodes use dhcp without static mapping. I have a dns server which can resolve local lookups. Would it be possible to continue using dhcp for nodes when the cluster seems to expect static mappings?

 

[fiona]

Proxmox VE nodes are expected to be configured with a static IP AFAIK.

 

[forresthopkinsa]

So are clusters explicitly not supported when using DHCP?

 

[fiona]

Hi,

So are clusters explicitly not supported when using DHCP?

AFAIK, you need static IPs for the Proxmox VE nodes. And the hostname should be resolvable via /etc/hosts. Maybe there's some trick out there to make it work with DHCP, but none that I know of.

See also: https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#sysadmin_network_configuration