0

I am having some difficulty with MAAS 2.3 enlisting servers. These are the steps I took to install (I have all Intel AMT servers in the lab, no IPMI). In my prior life I installed MAAS 1.9 using similar methods with no issues.

  • Installed 16.04 LTS for the MAAS server (two NICs, one onboard with PXE boot and static IP set as 10.10.1.100 with gateway 10.10.0.1 (10.10.0.0/16 network)).
  • Changed the network settings so the private network can reach to 10.71.101.1 (so I thought maybe?). OPT1 is internet facing (outbound):

    # The primary INTERNET (via OPT1) facing network interface.
    auto enp1s0
    iface enp1s0 inet static
        address 10.71.101.1
        netmask 255.255.0.0
        network 10.71.0.0
        broadcast 10.71.255.255
        gateway 10.71.0.1
        dns-nameservers 10.10.0.1 10.71.0.1
    
    # The private/MAAS facing network interface. Onboard NIC.
    auto eno1
    iface eno1 inet static
        address 10.10.0.1
        netmask 255.255.0.0
        network 10.10.0.0
        broadcast 10.10.255.255
        dns-nameservers 10.10.0.1 10.71.101.1
        post-up iptables -t nat -A POSTROUTING -o enp1s0 -j SNAT --to-source 10.71.101.1
        post-down iptables -t nat -D POSTROUTING -o enp1s0 -j SNAT --to-source 10.71.101.1
    
  • Installed etherwake and wsmancli for AMT support.

  • Installed MAAS 2.3. Enabled DHCP on 10.10.0.0/16 with suggested settings). Added 16.04 LTS image to be used in commissioning and SSH key etc on the MAAS web-ui.
  • For the nodes to be enlisted I have two NICs, the onboard NIC used for Intel AMT has a static address, like 10.10.1.100 PXE boot. The other NIC is also connected to the private network (10.10.0.0/16).
  • After turning on the machine, I cannot see any node enlisted.
  • I log on to the box being enlisted to see what is wrong. I can see a line in the cloud init output logs that 10.71.101.1 is not reachable during the node-enlist process. So it is... The node cannot get out to reach 10.71.101.1, the very same node that PXE boot is actually getting the boot image from.
  • When I checked the ifconfig on that particular node, I can see that the onboard NIC is given a new DHCP address from MAAS's DHCP pool for the private subnet. Not the static IP I designed for the onboard NIC. The second NIC does not have IPv4 address.

The 50-cloud-init.cfg in /etc/network/interfaces.d has this content

iface eno1 inet dhcp
    broadcast 10.10.255.255
    dns-nameservers 10.71.101.1
    dns-search maas.
    gateway 10.10.255.254

This perplexes me as the onboard nic already has the gateway defined as 10.10.0.1, which works for PXE boot. Why use this IP where the box cannot even access 10.10.255.254 to begin with?

Any help opening my eyes to see what I am doing wrong here would be greatly appreciated.

0

Alright... Looks like the issue was completely related to the gateway really. As MAAS is sending the gateway as part of the DHCP hand out, node being enlisted is given 10.10.255.254 IP as a gateway. I set that up by simply accepting the defaults without double checking.

Once I put the Gateway IP in the DHCP server as 10.10.0.1 all worked well, enlisted nodes finding the MAAS to be enlisted!

I hope this could help someone in he future.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.