Open Stack Dashboard Horizon configuration assistance

Asked by Richard Cunningham on 2013-03-14

Hello All,

I am in the process of installing Openstack on Ubuntu Server 12.04 and am up to the final stage to configure the horizon web front end. When I go to http://<myserver>/horizon and log in with my admin account (which is configured in Keystone), I get an Internal Server Error.

I am not sure where to start troubleshooting this, as their doesn't seem to be an openstack-dashboard error logs. Apache error log shows the following:
[Wed Mar 13 16:47:37 2013] [error] File "/usr/lib/python2.7/dist-packages/keystoneclient/service_catalog.py", line 62, in url_for
[Wed Mar 13 16:47:37 2013] [error] raise exceptions.EndpointNotFound('Endpoint not found.')

I did find an article where is talks about /etc/openstack-dashboard/local_settings.py using 'Member' as the default role from keystone. I have created this role to no avail.

All my nova services are up and showing as healthy:
Code:
Binary Host Zone Status State Updated_At
nova-consoleauth ms-os01 nova enabled :-) 2013-03-13 11:45:18
nova-cert ms-os01 nova enabled :-) 2013-03-13 11:45:18
nova-scheduler ms-os01 nova enabled :-) 2013-03-13 11:45:14
nova-compute ms-os01 nova enabled :-) 2013-03-13 11:45:15
nova-network ms-os01 nova enabled :-) 2013-03-13 11:45:15
Any ideas would be welcome

Thanks in advance,

Rich

Question information

Language:
English Edit question
Status:
Answered
For:
OpenStack Dashboard (Horizon) Edit question
Assignee:
No assignee Edit question
Last query:
2013-03-14
Last reply:
2013-03-14
Julie Pichon (jpichon) said : #1

(my bad, didn't notice this was posted twice and the other one was closed!)

Hi Richard.

What version of the Keystone client are you using? The endpoint not found may be inoffensive if you're using a version previous to 0.2.0 (see https://bugs.launchpad.net/python-keystoneclient/+bug/1036342).

Is there any other error traceback in the logs? Can you access another page directly e.g. /horizon/syspanel/users?

slice16 (paul-sanders87) said : #2

Hi Julie,

We currently have 2012.1+stable~20120824-a16a0ab9-0ubuntu2.5 installed.

We are able to go into other pages by adding the pages in a custom URL. The pages that get an internal server error are:

1_ Overview
2) Instances
3) Flavors
4) Quotas
5) Projects

Thanks

Paul

Julie Pichon (jpichon) said : #3

I think most of these pages require Nova. Can you paste the output of keystone endpoint-list? Could you check that the nova endpoint and keystone admin endpoint are reachable from the machine hosting Horizon? Also, does the instances page work for regular users?

What guide are you following to install OpenStack?

slice16 (paul-sanders87) said : #4

Thanks Julie, here is my endpoint-list:

+----------------------------------+-----------+--------------------------------------------------+--------------------------------------------------+---------------------------------------------+----------------------------------+
| id | region | publicurl | internalurl | adminurl | service_id |
+----------------------------------+-----------+--------------------------------------------------+--------------------------------------------------+---------------------------------------------+----------------------------------+
| 273a6bf087594f9fa827a48bea140ffd | RegionOne | http://192.168.10.200:8776/v1/%(tenant_id)s | http://192.168.10.200:8776/v1/%(tenant_id)s | http://192.168.10.200:8776/v1/%(tenant_id)s | c2b08b023a4145bda6201aafa3b30e91 |
| 3694d4084c7041969071f67543a5615f | RegionOne | http://192.168.10.200:8774/v2/%(tenant_id)s | http://192.168.10.200:8774/v2/%(tenant_id)s | http://192.168.10.200:8774/v2/%(tenant_id)s | 126821ee14c841198469fcd73aaba512 |
| 554eede7f16e4752ac0f89a7d3372aaa | RegionOne | http://192.168.10.200:5000/v2.0 | http://192.168.10.200:5000/v2.0 | http://192.168.10.200:35357/v2.0 | 5e5217bf351d43fca912084eee2cc835 |
| ab4a2fc8b14b469095a503eba1cfe92b | RegionOne | http://192.168.10.200:9292/v1 | http://192.168.10.200:9292/v1 | http://192.168.10.200:9292/v1 | c2b08b023a4145bda6201aafa3b30e91 |
| cdb97f0987f443a79ef80eb7093d7da7 | RegionOne | http://192.168.10.200:8888/v1/AUTH_%(tenant_id)s | http://192.168.10.200:8888/v1/AUTH_%(tenant_id)s | http://192.168.10.200:8888/v1 | 14de9dbc75aa4894aeb997a836f16b1f |
| ef74908242864dcea694e0f833735a8a | RegionOne | http://192.168.10.200:8773/services/Cloud | http://192.168.10.200:8773/services/Cloud | http://192.168.10.200:8773/services/Admin | 0f67acaa983743a1a3750ca19a0b9663 |
+----------------------------------+-----------+--------------------------------------------------+--------------------------------------------------+---------------------------------------------+----------------------------------+

All roles currently live on the same box, this is still a POC environment. It also affects standard member users.

I am using the guide at http://docs.openstack.org/trunk/openstack-compute/install/apt/openstack-install-guide-apt-trunk.pdf .

Thanks

Paul

Julie Pichon (jpichon) said : #5

Thank you for the extra information. I don't see anything obviously wrong with the endpoints themselves... Are there any errors in the nova logs? Can you run "nova list" successfully from the command line? "nova flavor-list"?

Setting DEBUG to True in the local_settings.py and restarting apache may also help in obtaining a more specific error message. I'm surprised the logs are empty when Horizon is returning an internal server error.

slice16 (paul-sanders87) said : #6

Turning on Debug and running the nova commands make more sense :)

I am now getting a Connection Refused error:

ConnectionRefused: '[Errno 111] Connection refused'

Request Method: GET
Request URL: http://192.168.10.200/horizon/syspanel/
Django Version: 1.4.1
Exception Type: ConnectionRefused
Exception Value:

ConnectionRefused: '[Errno 111] Connection refused'

Exception Location: /usr/lib/python2.7/dist-packages/novaclient/client.py in request, line 193
Python Executable: /usr/bin/python
Python Version: 2.7.3
Python Path:

['/usr/share/openstack-dashboard/openstack_dashboard/wsgi/../..',
 '/usr/lib/python2.7',
 '/usr/lib/python2.7/plat-linux2',
 '/usr/lib/python2.7/lib-tk',
 '/usr/lib/python2.7/lib-old',
 '/usr/lib/python2.7/lib-dynload',
 '/usr/local/lib/python2.7/dist-packages',
 '/usr/lib/python2.7/dist-packages',
 '/usr/share/openstack-dashboard/',
 '/usr/share/openstack-dashboard/openstack_dashboard']

Julie Pichon (jpichon) said : #7

Ok! So this is what usually happens when Nova is down or unreachable. Is Nova running at the port indicated in the endpoint list? Do the nova CLI commands respond successfully?

slice16 (paul-sanders87) said : #8

This is looking easier to troubleshoot now :)

Whenever I run a nova command I get a connection refused. I have just ran 'nova-manage logs errors' and can see the following errors:

Line 171 : 2013-03-14 12:56:00 1077 ERROR nova.openstack.common.rpc.common [-] AMQP server on 192.168.10.200:5672 is unreachable: [Errno 111] ECONNREFUSED. Trying again in 1 seconds.

So this is pointing to rebbitmq, but I know this works ok, as I can telnet onto 5672 from a number of different boxes.

Julie Pichon (jpichon) said : #9

Another question with a similar error was actually posted this morning: https://answers.launchpad.net/horizon/+question/224273 , perhaps the hints offered by the poster when closing the question will help you. This is getting into debugging Nova territory, you might be able to find better help by redirecting your question there!

slice16 (paul-sanders87) said : #10

Thanks for your help Julie, will do.

I will raise the question in nova :)

Can you help with this problem?

Provide an answer of your own, or ask Richard Cunningham for more information if necessary.

To post a message you must log in.