maas-devel team mailing list archive

[Diogo Matsubara <matsubara@xxxxxxxxxxxxx>]

Hi there,

today I tested maas - 0.1+bzr1268+dfsg-0ubuntu1~ppa1 from the experimental PPA.

My main focus was to try and get the multiple cluster controller
configuration working. I failed.

In the QA lab, machine lenovo-01 I have two VMs. One for the region
controller and another for the cluster controller.

The RC has 4 network interfaces:
eth0: 10.0.2.x used to connect to the VM LAN
eth1: 192.168.21.x the address of the RC pxe server (nodes 02-06 in
the lenovo lab are part of this subnet)
eth2: 192.162.22.x connected to the ipmi network for the lab
eth3: 10.98.0.x external ip address (API, web ui)

The CC has 3 network interfaces:
eth0: 10.0.2.x used to connect to the VM LAN
eth1: 192.168.20.x the cluster controller network (nodes 07-12 in the
lenovo lab are part of this subnet)
eth2: 10.98.0.x

In the RC:

apt-get install maas maas-dhcp maas-dns

dpkg-reconfigure maas-region-controller and point it to the PXE
address of the RC: 192.168.21.5

In the CC:

apt-get install maas-cluster-controller maas-dhcp

dpkg-reconfigure maas-cluster-controller and point it to the RC API
address: http://10.98.0.90/MAAS

Once this was done, the cluster controller shows up in the web ui and
I get this logs in CC's /var/log/maas/celery.log:
https://pastebin.canonical.com/76792/

The amqp address in the cluster controller is pointing to
192.168.21.5, but that's the PXE address of the RC. It should be
10.98.0.90, since rabbit is supposed to listen in the same ip address
of the API, correct? Or did I miss something during install?

I have also confirmed this bug
https://bugs.launchpad.net/ubuntu/+source/maas/+bug/1067558 (the
warning about maas-import-pxe-files doesn't disappear after you run
it) and filed a small UI glitch when you mouse over edit icons:
https://bugs.launchpad.net/bugs/1067925

Cheers,
-- 
Diogo M. Matsubara