CannotConfigureDHCP not handled

[Jeroen T. Vermeulen]

If configuration of the DHCPv6 server fails on the cluster controller, the exception bubbles back up into the region controller. (The region controller triggers the action through a signal when saving a cluster interface, and perhaps in other situations as well.)

The region controller does not handle this well. When the exception occurs during a UI request, it returns a bare “Internal server error” page.

[Jeroen T. Vermeulen (jtv)]

Reconfiguration of the DHCP server can trigger this either by actually failing, or by timing out.

[Julian Edwards (julian-edwards)]

This seems critical, can you comment jtv?

[Graham Binns (gmb)]

On 26 August 2014 15:01, Jeroen T. Vermeulen <email address hidden> wrote:
> The region controller does not handle this well. When the exception
> occurs during a UI request, it returns a bare “Internal server error”
> page.

It occurs to me (not for the first time) that we should have
middleware to at least give a nice OOPS/error page when this stuff
(i.e. anything unhandled) happens.

[Julian Edwards (julian-edwards)]

On Monday 13 Oct 2014 07:00:57 you wrote:
> On 26 August 2014 15:01, Jeroen T. Vermeulen <email address hidden>
wrote:
> > The region controller does not handle this well. When the exception
> > occurs during a UI request, it returns a bare “Internal server error”
> > page.
>
> It occurs to me (not for the first time) that we should have
> middleware to at least give a nice OOPS/error page when this stuff
> (i.e. anything unhandled) happens.

Yes. I'm sure I mentioned this on IRC the other day too :)

[Graham Binns (gmb)]

On 13 October 2014 08:23, Julian Edwards <email address hidden> wrote:
> Yes. I'm sure I mentioned this on IRC the other day too :)

You did :).

[Christian Reis (kiko)]

I've just run into a user on #maas who is facing this; his traceback is at http://pastebin.com/8FEcA2ta

I'm not sure why the failure is happening, though.

[Christian Reis (kiko)]

See also bug 1395896