[openstack-dev] [horizon][heat] Unusable error messages in dashboard for Orchestration

Fox, Kevin M Kevin.Fox at pnnl.gov
Fri Apr 11 15:21:59 UTC 2014


I've seen unusable error messages out of heat as well. I've been telling users (our ops guys) to look at the heat-engine logs when it happens and usually its fairly apparent what is wrong with their templates.

In the future, Should I report each of these I see as a new bug or add each to the existing bug?

Thanks,
Kevin
________________________________________
From: Steven Hardy [shardy at redhat.com]
Sent: Friday, April 11, 2014 3:24 AM
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [horizon][heat] Unusable error messages in dashboard for Orchestration

Hi Tom,

On Fri, Apr 11, 2014 at 01:05:00PM +0800, Tom Fifield wrote:
> Hi,
>
> Lodged a bug the day after Havana came out to hope to get this
> usability problem addressed.
>
> https://bugs.launchpad.net/horizon/+bug/1241395
>
> Essentially, if something goes wrong creating a stack through the
> dashboard, all the user ever sees is:
>
> "Stack creation failed."
>
>
> ... which is, a little less than useful in terms of enabling them to
> fix the problem.
>
> Testing using RC1 today, there is no improvement, and I was shocked
> to discover the the bug submitted was not even triaged during
> icehouse!
>
> Any ideas? :)

Thanks for drawing attention to this problem, unfortunately it looks like
this bug has fallen through the cracks so sorry about that.

I am a little confused however, because when I create a stack which fails,
it does provide fedback regarding the reason for failure in both the
topology and stack overview pages.

Heres the steps I went through to test:

1. login as demo user, select demo tenant
2. select Orchestration->Stacks page
3. Click Launch Stack, do direct input of this template:

heat_template_version: 2013-05-23

resources:
    wait_handle:
        type: OS::Heat::UpdateWaitConditionHandle

    wait_condition:
        type: AWS::CloudFormation::WaitCondition
        properties:
            Count: 1
            Handle: { get_resource: wait_handle }
            Timeout: 10

4. Click next, enter stack name "test", password "dummy" and Click "Launch"
5. The UI returns to the stack list view, with the test stack status "in
progress"
6. The stack creation fails, status moves to "failed"
7. Click on the stack name, takes me to "Stack Overview" page
8. Observe Status for failure:

Status
  Create_Failed: Resource CREATE failed: WaitConditionTimeout: 0 of 1 received


So there evidently are some other error paths which do not provide feedback
(when we fail to attempt stack creation at all) - the only way I could
reproduce your issue was by attempting to create a stack with a duplicate
name, which as you say does result in an uhelpful popup saying "Error:
Stack creation failed."

Comparing that with, for example, creating a stack with an invalid name
"123", which gives you a nice explanation of the problem on the form.  SO I
guess what you're seeing is a default response to some error paths which
are not explicitly handled.

What I think may help is if you can update the bug with steps to reproduce
the problem.  It may be there are several failure scenarios we need to
resolve, so getting specific details to reproduce is the first step.

Thanks!

Steve

_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev



More information about the OpenStack-dev mailing list