juju bootstrap ssh no route

Question

the problem is based on the context of VirtualBox, they are set on 3 virtual machines whose nodes are associated with a cluster controller, however when it completes the execution of "juju bootstrap" this ends with the following result:

Attempting to connect to node02-080027EB54F9.maas:22
Attempting to connect to 10.0.2.22:22

ERROR bootstrap failed: waited for 30m0s without being able to connect: ssh: connect to host 10.0.2.22 port 22: No route to host

Note that the node is set to the point that asks "login", but to make "ping" the machine is not responding.

I hope you can help me,

Best Regards

**Virtual Machine Configuration ** 

Cluster Controller
    ram: 2048 MB
    disk: 80,00 GB
    if1: 08:00:27:9E:54:A6 - 10.0.2.10 (Nat Network)
    if2: 08:00:27:22:F4:D6 - 192.168.1.37 (Bridge)

Node1
    ram: 512 MB
    disk: 8,00 GB
    if1: 08:00:27:92:36:A5 - 10.0.2.20 (Nat Network)

Node2
    ram: 512 MB
    disk: 8,00 GB
    if1: 08:00:27:EB:54:F9 - 10.0.2.22 (Nat Network)

Answer 1

Seems like a network problem to me. Are your VBox networks configured correctly? If your are installing a clean ubuntu on one of your node vms, is it reachable from the cluster controller? How are your network settings for maas? Is it a maas controlled network using dhcp or did your configured your preseed to use a static ip network with predefined ips? If you are using dhcp you should check the /var/log/syslog if you can find some dhcp requests from the bootstrapped server (cat /var/log/syslog | grep 'mac_of_bootstrapped_node' on cluster controller)

Answer 2

Did you setup your ssh key? If not instructions are on this page.

[https://maas.ubuntu.com/docs/juju-quick-start.html]

Another problem could be dns not resolving correctly or at all. From your maas regional controller try and ping other computers on the network. If it cannot resolve any hostnames then you have a problem with bind9.