Internal Server Error when trying to deploy a node when maas-dhcp was not running

[Andres Rodriguez]

When trying to deploy a machine, MAAS showed an internal server error. The reason seems to be due to the fact that maas-dhcpd was not running. clusterd.log showed the following issue:

2015-02-17 11:52:20-0500 [ClusterClient,client] Cluster '83a8490d-a5eb-4748-95fd-c6e9b8a02951' registered (via unleashed:pid=4615).
dhcpctl_connect: no more
dhcpctl_connect: no more
dhcpctl_connect: no more
dhcpctl_connect: no more
2015-02-17 11:57:44-0500 [

[Andres Rodriguez (andreserl)]

==> /var/log/maas/maas.log <==
Feb 20 16:22:24 unleashed maas.dhcp: [ERROR] Could not create host map for 52:54:00:6f:08:68 with address 192.168.122.151: Command `omshell` returned non-zero exit status 0:#012> > > > not connected.#012> no open object.#012> no open object.#012> no open object.#012> no open object.#012> not connected.#012>

[Andres Rodriguez (andreserl)]

maybe a notification should be shown telling the user that the cluster controller the node is connected to does not have maas-dhcpd running.

[Andres Rodriguez (andreserl)]

this also applies when trying to deploy via the bulk actions. The only difference is that the bulk actions do not show this internal server error.

[Shades (initialhit)]

Seeing this too:

ERROR] Could not remove host map for 10.0.0.128: Command `omshell` returned non-zero exit status 0:#012> > > > not connected.#012> no open object.#012> not connected.#012> not connected.#012>

Just ran updates as well. Shows up with bulk or individual nodes. Only happened after I turned off DHCP on one of my interfaces

Changes I did to make this happen:
Network 1: DHCP -> Unmanaged
Network 2: DHCP / DNS -> DHCP / DNS

Let me know if I can help provide more details :) If I have time I will look into what code is causing this.

Thanks,

Shades

[Shades (initialhit)]

Deleting all nodes and then turning off dhcp and then re-adding nodes works. Luckily for me I hadn't fully built out my openstack network yet.