← Back to team overview

openstack team mailing list archive

Re: Horizon - Internal Server Error when hitting /nova/instances_and_volumes/

 

On May 7, 2013, at 2:01 PM, Daniel Ellison <daniel@xxxxxxxxxx> wrote:
> As the subject says, I'm having issues getting at "Instances & Volumes" (also "Images & Snapshots") in Horizon. I'm running grizzly on precise. Everything else works fine; the entire "Admin" tab works as expected. The "Overview" and "Access & Security" also work fine.
> 
> Is there any way to see what calls are being made from horizon to nova? I debugged some previous issues by using --debug on some python client calls. But I don't think there's an equivalent in this case.
> 
> The Apache log on my horizon machine (a VM under nova) shows the 500 error, then has a bunch of 404s when trying to retrieve media and other resources, e.g. /nova/images_and_snapshots/dashboard/css/style.css. For all other calls there are no 404 errors (and no 500 errors, needless to say). 
> 
> I'm only using Nova, Keystone and Glance for the moment. So no Cinder to consider, and no attached volumes. Does any of this sound familiar to anyone?

I think I know the source of the problem. I first had to add a logging section for "openstack_dashboard" in local_settings.py despite the fact that there was already a section for "horizon". Then I was able to see the exception that was happening:

    ServiceCatalogException: Invalid service catalog service: volume

I don't have Cinder installed, but it looks like Horizon is expecting it to be there. I don't have a volume endpoint, as far as I can tell. "nova endpoints" only lists glance nova, ec2 and keystone.

Is there a way to disable this behaviour in Horizon? Or is there a setting I'm missing?

Thanks,
Daniel

Follow ups

References