Ask Your Question

Heather's profile - activity

2016-10-03 07:42:41 -0500 received badge  Notable Question (source)
2016-09-26 22:52:04 -0500 received badge  Popular Question (source)
2016-09-24 12:10:00 -0500 commented answer Is it necessary to have ceilometer running to obtain the status of a stack during creation?

The scenario I am working with is someone is polling for status of the stack. Would the stack status still be available fifteen seconds after the heat stack either completed successfully or if it failed without ceilometer enabled? I realize that if you are waiting in a synchronous fashion for the

2016-09-24 12:09:14 -0500 answered a question Is it necessary to have ceilometer running to obtain the status of a stack during creation?

The scenario I am working with is someone is polling for status of the stack. Would the stack status still be available fifteen seconds after the heat stack either completed successfully or if it failed without ceilometer enabled? I realize that if you are waiting in a synchronous fashion for the api call outcome heat will produce it. For whatever reason the group I am assisting does not seem to be doing that. I am trying to determine if ceilometer is needed to determine stack status post creation (or failure). Basically, does such status depend upon the capture of lifecycle events in ceilometer. And thank you for your reply.

2016-09-23 08:21:51 -0500 asked a question Is it necessary to have ceilometer running to obtain the status of a stack during creation?

I am working in an environment where ceilometer has been turned down. I was wondering what impact that had on the capture of lifecycle events, and if one could still get an accurate status from heat during the process of stack creation?

2016-09-20 12:06:57 -0500 marked best answer Need to get ip of a port in HOT template

I've tried numerous methods to pull in the IP address from a port on a vm without success.

Here is a snippet of the HOT doc.

The bolded part of the code is part that is not working. If I put a literal in there it will work fine. I'm new to Python so there may be some nuance of that I'm not understanding.

The spec for get_attr is:

get_attr: - <resource id=""> - <attribute name=""> - <key index="" 1=""> (optional) - <key index="" 2=""> (optional) - ...

Is the resource id the resource name in this case? wmwcvm1_port or is it the actual uuid of the resource? If it is the uuid can I use get_resource to pull that in because I won't know that value at run time?

I'm assuming the attribute name to be fixed_ips. I wrote some python code to pull the fixed_ip's out to take a look. This is what the API returns on a neutron list_ports() call:

fixed_ips : [{u'subnet_id': u'd4714185-cab2-4c5a-9d91-69a63f5dc063', u'ip_address': u'192.168.3.2'}]

  #############################################################
  # Define the VM and attach to the network
  #############################################################
  wmwcvm1:
    type: OS::Nova::Server
    metadata: http://haasaet.web.att.com/depot/deploy-chef-client.sh
    properties:
      name: wmwcvm1
      key_name: widgetco
      image: { get_param: image_id }
      flavor: { get_param: instance_type }
      networks:
        - port: { get_resource: wmwcvm1_port }
      user_data:
        str_replace:
          template: { get_file:  "http://haasaet.web.att.com/depot/hot-skel.sh" }
          params:
            $APP_HOST_IP_ENV: **{ get_attr: [ wmwcvm1_port, fixed_ips, ip_address, 0 ] }**
            $HOTSKEL2: { get_file:  "http://haasaet.web.att.com/depot/milestone2-run-recipe.sh" }

  wmwcvm1_port:
    type: OS::Neutron::Port
    properties:
      network_id: { get_param: zone1_id }
      fixed_ips:
        - subnet_id: { get_param: zone1_subnet_id }
      security_groups: [{ get_resource: server_security_group }]
2016-08-24 21:58:25 -0500 received badge  Famous Question (source)
2016-08-24 21:58:25 -0500 received badge  Notable Question (source)
2016-02-09 20:51:31 -0500 received badge  Notable Question (source)
2015-11-06 14:48:30 -0500 commented question Does Fuel absolutely require root to run?

It is against our security organization's policy to do anything you don't absolutely have to (i.e. initial installs and setup of the OS) as true root. sudo allows for audit trails which are required. Bypassing that is an issue for us.

2015-11-03 14:43:01 -0500 received badge  Popular Question (source)
2015-10-30 11:21:10 -0500 asked a question Does Fuel absolutely require root to run?

I'm being told my Mirantis that Fuel has to run as true root. Not sudo. That concerns me. Is that really true?

2015-09-04 17:23:03 -0500 received badge  Autobiographer
2015-06-18 04:03:51 -0500 received badge  Famous Question (source)
2015-05-12 11:01:01 -0500 received badge  Famous Question (source)
2015-04-15 02:05:47 -0500 marked best answer Can you reverse generate a HOT template from an existing stack?

I see numerous API calls that could be used to piece this together, but I was wondering if something already existed. I see the heat resource-template command that generates a skeleton HOT doc. heat resource-list, resource-show which provide some of the data.

it would be nice to be able to re-baseline your stack(s) to account for situations where people might use the dashboard or direct api calls to implement changes without your knowledge.

Update

This is excellent information. Thank you. The question we were grappling with was how to manage the change process and what to establish as the source document for the implementations.

2015-04-15 01:44:57 -0500 received badge  Famous Question (source)
2015-04-10 14:20:01 -0500 marked best answer Type not found - OS::Heat::SoftwareDeployments

According to the docs it was available in 2014.1 (Icehouse).

My RHEL RPMs are 2014.1.3-1.e16 for openstack-heat-engine.noarch (same for common).

openstack-heat-templates.noarch is 0-0.3.20140407git.e16

Should this type be in this deployment? If so, where might I start looking for why I can't seem to find it? If not, what version do I need?

Thanks.

2015-03-30 02:12:13 -0500 received badge  Taxonomist
2015-02-22 12:35:10 -0500 received badge  Notable Question (source)
2015-02-08 02:30:37 -0500 received badge  Self-Learner (source)
2015-02-08 02:30:11 -0500 received badge  Famous Question (source)
2015-02-08 02:30:11 -0500 received badge  Famous Question (source)
2015-02-02 16:12:18 -0500 received badge  Famous Question (source)
2015-02-02 15:28:47 -0500 received badge  Popular Question (source)
2015-02-02 15:28:47 -0500 received badge  Notable Question (source)
2015-01-15 13:08:13 -0500 marked best answer What is the proper format to pass allocation_pools to a nested stack?

I have tried numerous combinations trying to get this to work. I have tried both comma_delimited_list and json types.

When I hardcode this into the nested stack:

allocation_pools: [ { "start": "192.168.20.2", "end": "192.168.20.254" } ]

it works.

When I try to pass it in as a parameter I get an error that says it is not a list.

This is a section of the resource def where it is passed in:

 NestZone2:
    type: zone_network.template
    properties:
      zone_name: nzone2
      cidr: 192.168.21.0/24
      allocation_pools: [ { "start": "192.168.21.2", "end": "192.168.21.254"
} ]
      gateway_ip: 192.168.21.1
      shared_network: 'False's
      enable_dhcp: 'True'

In the nested template the parameter is defined as:

  allocation_pools:
    type: json
    description: Network Allocation Pools

And the resource snippet in the nested template is

  shared_network:
    properties:
      name: |
        str_replace:
          template: |
            $subnet
          params:
            $subnet: { get_param: zone_name }-subnet
      network_id:
        get_resource: zone_network
      allocation_pools:
        get_param: allocation_pools
      cidr:
        get_param: cidr

I've tried various yaml formats for a map in a list without success. I either get the error that it is not a list or that it is not a map. I figure this is a duh question, but I'm tired of beating my head against it.

2015-01-15 03:58:26 -0500 received badge  Popular Question (source)
2015-01-15 03:58:26 -0500 received badge  Notable Question (source)
2014-12-18 08:08:11 -0500 received badge  Famous Question (source)
2014-12-08 11:20:27 -0500 received badge  Notable Question (source)
2014-12-04 03:43:52 -0500 received badge  Notable Question (source)
2014-11-25 16:40:55 -0500 received badge  Popular Question (source)
2014-11-25 16:40:55 -0500 received badge  Notable Question (source)
2014-11-25 16:40:55 -0500 received badge  Famous Question (source)
2014-11-21 09:46:11 -0500 asked a question Can Keystone roles be tied to stack parameters?

I know this would be a future potential enhancement. I can certainly see a need for adding a property to a parameter that defines which keystone role can modify that parameter. When dealing with less than experienced personnel there are some parameters you may want an admin to be able to modify while others should be left to more experienced architects.

2014-11-18 00:09:11 -0500 received badge  Popular Question (source)
2014-11-17 16:33:53 -0500 commented answer Can you use autoscaling in combination with shelving and unshelving instances?

Thanks. Long application install times would make something like that very handy.

2014-11-17 16:33:53 -0500 received badge  Commentator
2014-11-17 16:01:23 -0500 received badge  Popular Question (source)
2014-11-17 14:29:58 -0500 commented answer Does SoftwareConfig and SoftwareDeployment work out of the box?

Getting the image installed with the required diskimage-builder has proven challenging and still not working yet. The documentation is almost non-existent.

2014-11-17 11:47:47 -0500 asked a question Intermittent issues with Cinder block storage attaching.

We are seeing intermittent issues where we allocate and attach cinder storage in HOT where it works fine sometimes and other times the storage does not show up.

We attach it to /dev/vdb. Most times it's there and works great. Occasionally, it does not. You do a df and it doesn't show. Do a soft reboot of the instance and it will then show up. We are doing steps serially to prevent race conditions and our chef recipe will loop for 2 minutes waiting for it to become available. We cannot re-create this at will. As I said, most of the time it works.

Is this a known issue?

2014-11-17 11:42:02 -0500 asked a question Can you use autoscaling in combination with shelving and unshelving instances?

Rather than having to completely rebuild an instance it seems like shelving and unshelving an already built instance might be a faster way to get a resource available. Just a thought.

2014-11-15 04:13:15 -0500 received badge  Notable Question (source)
2014-11-04 21:37:04 -0500 received badge  Notable Question (source)
2014-10-31 13:16:34 -0500 received badge  Popular Question (source)
2014-10-31 10:31:08 -0500 received badge  Famous Question (source)
2014-10-30 16:48:06 -0500 answered a question Does SoftwareConfig and SoftwareDeployment work out of the box?

The answer appears to be no.

Refer to the end of this deck by Steve Baker (RedHat).

deck

I am going to try and get those built into our images tomorrow and re-try.

2014-10-30 10:12:44 -0500 asked a question Why doesn't SoftwareConfig and SoftwareDeployment show up in my stack resource list?

I have a SoftwareConfig, CloudConfig, MultipartMime, and SoftwareDeployment resource that seems to be being ignored at stack build time and they don't show up in the resource list.

2014-10-28 12:58:17 -0500 asked a question Does SoftwareConfig and SoftwareDeployment work out of the box?

Or do you have to do something special with the hook scripts first? Can you point me at a decent example?