* Move the commision_vm/node to functions.sh
* if machine already exists in maas, then re-use it
* move node variables above the node_num, to calculate correctly
* parametrise maas_kernel_opts and maas_name, so that it's not specific
for an environment
* update boot-source-selections such that we have both focal and bionic
by default
* Add comments to the default.config file
* move the node_count to the default.config, this will allow to use
customised value if needed
* If pod-console-logging tag is to be added, then add correctly,
Otherwise, if we are adding fresh pods after tag creation, the pods
will fail to add. This also allows for other cases to be fulfilled if
required
* enable/disable managed on the subnets. This resolves the problem where AUTO IP
doesn't seem to work. This is a workaround to help with this
* Revert the AUTO mode, now that the above issue is resolved
* More tabs to spaces, so that formatting and disaply looks the same
on terminal and text editor
* Ensure that only amd64 are being sync'd
* If using MAAS 2.9, then focal gets downloaded by default, so add
bionic to the list too
* Cross out items in TODO that have now been implemented
* Make password and virsh user consistent on hp01
* Update the config to reflect the network config from hp01
* Add timezone and byobu config in cloud init
* This will allow to run a quick script to wipe the headers
of the disk, so that we don't have to login after a
re-comission of the machine to manually reboot it
* Update script to add all stuff from install_kvm into it, rather than
runcmd.
* Update the virsh user as per the MAAS code
* Add support to add focal, only change of packages
* Add pod creation, so that when we remove the hypervisor
then all the VMs also get removed automatically
* virtinst is only used when creating/installing VMs, so
only install on hypervisor nodes
* Only grab the top vlan if we have 2, this typically
means that we have an ipv4 and ipv6 IP. Could probably
do this better, but keep it simple for the time being
* Update the function, such that when subnet_id is null, then based on
the vlan, grab the details from the subnets defined, and update the
links based on this. This ensures that all subnets are auto-assigned
and the externel_vlan is dhcp from the upstream network
* Revert the network_auto function, as the above change would help with
the issues we faced before
* Update all the maas function calls, and assign them to a variable or
direct the output to /dev/null, as we don't want the screen to be
filled with all the information
TODO: Have viable output to the CLI, so that you have some details of
what is going on
* So that all the networks now will have be either static, auto or dhcp
* network_auto will not work out of the box, it can only effectively
work after commissioning of a machine
* Move stuff from common to default, as it didn't make sense
* use virsh_user instead, that was its more customisable
* Revert the change to manage-maas-nodes.sh now use extra function to
call maas_auto_assign_networks instead
* More comments to the code
