I am facing issues with juju initiated lxd containers in all deployed machines.
There are two bridges br-ens160(juju created) and lxdbr0(default lxd)
From br-ens0:
The containers are taking DHCP IPs of the VLAN where juju and maas reside.
I can ping the containers from the parent host and vice-versa but I can’t ping them from any other host on the VLAN nor the containers can ping other hosts and the gateway.
Due to this the juju machine state of the containers is still pending.
From lxdbr0:
I can ping the gateway as well as other hosts on the network. This one has a different subnet.
While trying with juju model-config container-networking-method=local i get
You definitely want to keep using the default container-networking-method=provider.
We would need more information regarding the network set-up here.
I am able to launch LXD instances on MAAS-deployed machines.
Using Juju 2.9 and MAAS 2.8.6:
juju status
Machine State DNS Inst id Series AZ Message
0 started 172.16.99.3 m-03 focal zone1 Deployed
0/lxd/0 started 172.16.99.4 juju-6011c4-0-lxd-0 focal zone1 Container started
Machine 0 interfaces.
5: br-ens4: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
link/ether 52:54:00:14:1c:22 brd ff:ff:ff:ff:ff:ff
inet 172.16.99.3/24 brd 172.16.99.255 scope global br-ens4
valid_lft forever preferred_lft forever
inet6 fe80::5054:ff:fe14:1c22/64 scope link
valid_lft forever preferred_lft forever
7: 0lxd0-0@if6: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br-ens4 state UP group default qlen 1000
link/ether c2:78:c7:ce:db:6c brd ff:ff:ff:ff:ff:ff link-netnsid 0
Machine 0 routes.
default via 172.16.99.254 dev br-ens4 proto static
10.134.70.0/24 dev lxdbr0 proto kernel scope link src 10.134.70.1 linkdown
172.16.99.0/24 dev br-ens4 proto kernel scope link src 172.16.99.3
The subnet deployed to has MAAS-managed DHCP. When I look at the subnet listing in MAAS, I can see the assigned address.
