Validations

Validations

Since the Newton release, TripleO ships with extensible checks for verifying the undercloud configuration, hardware setup, and the deployment to find common issues early.

The TripleO UI runs the validations automatically. While work is underway for having them in the command line workflow, too, they need to be run by calling a Mistral workflow currently.

The validations are assigned into various groups that indicate when in the deployment workflow are they expected to run:

  • prep validations check the hardware configuration of the undercloud node and should be run before openstack undercloud install.
  • pre-introspection should be run before we introspect nodes using Ironic Inspector.
  • pre-deployment validations should be executed before openstack overcloud deploy
  • post-deployment should be run after the overcloud deployment has finished.
  • pre-upgrade try to validate your undercloud before you upgrade it.

Note that for most of these validations, a failure does not mean that you’ll be unable to deploy or run OpenStack. But it can indicate potential issues with long-term or production setups. If you’re running an environment for developing or testing TripleO, it’s okay that some validations fail. In a production setup they should not.

The list of all existing validations and their groups is on the tripleo-validations documentation page.

Running a single validation

If you want to run one validation in particular (because you’re trying things out or want to see whether you fixed the failure), you can run it like so:

$ source stackrc
$ openstack workflow execution create tripleo.validations.v1.run_validation '{"validation_name": "undercloud-ram"}'

This will run the “undercloud-ram.yaml” validation.

The undercloud comes with a set of default validations, which are stored in a Swift container called tripleo-validations. To get the list of available validations, run openstack container list tripleo-validations. To get the name of the validation (undercloud-ram here) find the right one there and use its filename without an extension.

The command will return an execution ID, which you can query for the results. To find out whether the validation finished, run:

$ openstack workflow execution show ID

Note the State value: it can be either RUNNING or SUCCESS. Note that success here only means that the validation finished, not that it succeeded. To find that out, we need to read its output:

$ mistral execution-get-output ID | jq .stdout -r

Note

There is an openstack workflow execution show output command which should do the same thing, but it currently doesn’t work in all environments supported by TripleO.

This will return the hosts the validation ran against, any warnings and error messages it encountered along the way as well as an overall summary.

Custom validations

Support for custom validations has been added in the Rocky development cycle. It allows operators to add their own bespoke validations, in cases when it’s not appropriate to include these in the set of default TripleO validations.

Custom validations are associated with deployment plans and stored in the plan’s Swift container, together with the rest of the plan files.

To add custom validations for a deployment plan, create a custom-validations subdirectory within the deployment plan Swift container and place the validations yaml files there.

To run a custom validation, follow the same procedure as for running one of the default validations - determine the name of the validation by listing the contents of the custom-validations subdirectory, and supply that name (without the .yaml extension) to the run_validation workflow.

If a validation with the same name is found both in the set of default validations, and in custom validations, the custom validation is always picked. This means that, if you wish to override a default validation with your custom implementation of it, all you need to do is create a validation with the same name and place it in the custom-validations subdirectory of the Swift container holding the deployment plan.

Running a group of validations

The deployment documentation highlights places where you can run a specific group of validations. Here’s how to run the pre-deployment validations:

$ source stackrc
$ openstack workflow execution create tripleo.validations.v1.run_groups '{"group_names": ["pre-deployment"]}'

Unfortunately, getting the results of all these validations is more complicated than in the case a single one. You need to see the tasks the workflow above generated and see their results one by one:

$ openstack task execution list

Look for tasks with the run_validation name. Then take the ID of each of those and run:

$ mistral task-get-result ID | jq .stdout -r

Note

There is a task execution show result command that should do the same thing, but it’s not working on all platforms supported by TripleO.

Since this can be tedious and hard to script, you can get the list of validations belonging to a group instead and run them one-by-one using the method above:

$ openstack action execution run tripleo.validations.list_validations '{"groups": ["pre-deployment"]}' | jq ".result[] | .id"

Another example are the “pre-upgrade” validations which are added during the P development cycle (tracked with this blueprint). These can be executed as the example above but instead using the “pre-upgrade” group:

openstack workflow execution create tripleo.validations.v1.run_groups '{"group_names": ["pre-upgrade"]}'

+-------------------+--------------------------------------+
| Field             | Value                                |
+-------------------+--------------------------------------+
| ID                | 3f94a17b-835b-4a82-93af-a6cddd676ed8 |
| Workflow ID       | e211099f-2c9b-46cd-a536-e38595ae8e7f |
| Workflow name     | tripleo.validations.v1.run_groups    |
| Description       |                                      |
| Task Execution ID | <none>                               |
| State             | RUNNING                              |
| State info        | None                                 |
| Created at        | 2017-06-29 12:01:35                  |
| Updated at        | 2017-06-29 12:01:35                  |
+-------------------+--------------------------------------+

You can monitor the progress of the execution by getting its status and also output:

mistral execution-get $ID
mistral execution-get-output $ID

When any of the validations fail the execution will have a ERROR status. You can query the individual validations in the group to determine the exact reasons that the validation fails. For example:

    for i in $(mistral execution-list | grep tripleo.validations.*ERROR | awk '{print $2}'); do mistral execution-get-output $i; done
{
    "result": "Failure caused by error in tasks: get_servers\n\n  get_servers [task_ex_id=a6ef7d32-4678-4a58-85fe-bf2da8a963ae] -> Failed to run action [action_ex_id=3a9a81e2-d6b0-4380-8985-41d6f4e18f3a, action_cls='<class 'mistral.actions.action_factory.NovaAction'>', attributes='{u'client_method_name': u'servers.list'}', params='{}']\n NovaAction.servers.list failed: <class 'keystoneauth1.exceptions.connection.ConnectFailure'>: Unable to establish connection to http://192.168.24.1:8774/v2.1/servers/detail: ('Connection aborted.', BadStatusLine(\"''\",))\n    [action_ex_id=3a9a81e2-d6b0-4380-8985-41d6f4e18f3a, idx=0]: Failed to run action [action_ex_id=3a9a81e2-d6b0-4380-8985-41d6f4e18f3a, action_cls='<class 'mistral.actions.action_factory.NovaAction'>', attributes='{u'client_method_name': u'servers.list'}', params='{}']\n NovaAction.servers.list failed: <class 'keystoneauth1.exceptions.connection.ConnectFailure'>: Unable to establish connection to http://192.168.24.1:8774/v2.1/servers/detail: ('Connection aborted.', BadStatusLine(\"''\",))\n"
}
{
    "status": "FAILED",
    "result": null,
    "stderr": "",
    "stdout": "Task 'Fail if services were not running' failed:\nHost: localhost\nMessage: One of the undercloud services was not active. Please check openstack-heat-api first and then confirm the status of undercloud services in general before attempting to update or upgrade the environment.\n\nTask 'Fail if services were not running' failed:\nHost: localhost\nMessage: One of the undercloud services was not active. Please check openstack-ironic-api first and then confirm the status of undercloud services in general before attempting to update or upgrade the environment.\n\nTask 'Fail if services were not running' failed:\nHost: localhost\nMessage: One of the undercloud services was not active. Please check openstack-zaqar first and then confirm the status of undercloud services in general before attempting to update or upgrade the environment.\n\nTask 'Fail if services were not running' failed:\nHost: localhost\nMessage: One of the undercloud services was not active. Please check openstack-glance-api first and then confirm the status of undercloud services in general before attempting to update or upgrade the environment.\n\nTask 'Fail if services were not running' failed:\nHost: localhost\nMessage: One of the undercloud services was not active. Please check openstack-glance-api first and then confirm the status of undercloud services in general before attempting to update or upgrade the environment.\n\nFailure! The validation failed for all hosts:\n* localhost\n"
}
Creative Commons Attribution 3.0 License

Except where otherwise noted, this document is licensed under Creative Commons Attribution 3.0 License. See all OpenStack Legal Documents.