Cyclades api error: Server takes too long to load

[Collins Udanor]

Hello,

I am done setting up cyclades, but when I start it, it takes so long a time to load if it does not throw API error: "It seems the server takes too long to respond, please check your network connectivity"

If it does load completely and I want to create a new VM, it throws the same error again. Once the dialog box pops up it blocks me from creating the VM. I am able to see the images in pithos from cyclades new VM dialog.

Someone please assist.

Thanks 

Collins

[Brian Candler]

That error message may be misleading.  As I understand it, Cyclades communicates via a messaging system (rabbitmq).  So "Server takes too long to respond" could simply mean that the server is not running, or the messaging system has a problem.

Have you looked in any logs? Is the cyclades service running? Is rabbitmq running?

[Brian Candler]

When you get problems with a client side application, the best thing to do is to run the developer console in your browser. I'm using chrome.  It gives very clear error messages.

Using this I was able to find the following error:

Request URL: https://node1.example.com/node1.example.com/astakos/ui/get_services?callback=jQuery17206833499549575459_1489757485953&_=1489757487683

Request Method:GET

Status Code:404 NOT FOUND

You can see the Request URL is obviously broken. So did

grep -R node1.example.com /etc

to look at all the embedded URLs. And the problem was here:

/etc/synnefo/20-snf-cyclades-app-cloudbar.conf:CLOUDBAR_SERVICES_URL = 'https:/node1.example.com/astakos/ui/get_services'

There should be two slashes after https, not one. I fixed that and restarted gunicorn.

After that, still getting timeout errors. The URLs are:

https://node1.example.com/cyclades/image/v1.0/images/detail?_=1489757806841

https://node1.example.com/cyclades/compute/v2.0/flavors/detail?_=1489757806852

https://node1.example.com/cyclades/_astakos/account/resources?_=1489757806855

and the developer console showed these fetches were aborted after 10 seconds (hence your "too long to respond" errors). So there's some other problem in your system as well.

[Brian Candler]

FYI, there was a DNS problem:

  File "/usr/lib/python2.7/dist-packages/astakosclient/utils.py", line 43, in decorator

    raise err

ConnectionError: ('[Errno 3] name does not exist', '', 500)

2017-03-17 14:13:03,487 [7518] synnefo.api __init__ [ERROR] Failed to send request: DNSError(3, 'name

does not exist')

And there was also a rabbitmq problem (account hadn't been created properly).

Moral is: you have to *really* follow the instructions carefully. There are lots of moving parts. All of them must be correct, so you'd better understand each step and be sure everything is correct. You don't always get good error reports if something isn't done properly.

For example, the rabbitmq problem was reported in /var/log/synnefo/dispatcher.log, but it could only tell that the connection was dropped:

2017-03-15 12:57:30,808 amqp amqp_puka [ERROR] Cannot connect to host node1.example.com:5672: [Errno 104] Connection reset by peer

You had to look in /var/log/rabbitmq/rabbit\@node1.log to find the problem:

{channel0_error,starting,

                {amqp_error,access_refused,

                            "PLAIN login refused: user 'synnefo' - invalid credentials",

                            'connection.start_ok'}}