Profiles¶
Concept¶
A Profile is the mould used for creating a Node to be managed
by the Senlin service. It can be treated as an instance of a
Profile Type with an unique ID. A profile encodes the information
needed for node creation in a property named spec.
The primary job for a profile type implementation is to translate user provided JSON data structure into information that can be consumed by a driver. A driver will create/delete/update a physical object based on the information provided.
Listing Profiles¶
To examine the list of profile objects supported by the Senlin engine, you can use the following command:
$ openstack cluster profile list
+----------+----------+--------------------+---------------------+
| id | name | type | created_at |
+----------+----------+--------------------+---------------------+
| 560a8f9d | myserver | os.nova.server-1.0 | 2015-05-05T13:26:00 |
| ceda64bd | mystack | os.heat.stack-1.0 | 2015-05-05T13:26:25 |
| 9b127538 | pstack | os.heat.stack-1.0 | 2015-06-25T12:59:01 |
+----------+----------+--------------------+---------------------+
Note that the first column in the output table is a short ID of a profile
object. Senlin command line use short IDs to save real estate on screen so
that more useful information can be shown on a single line. To show the full
ID in the list, you can add the --full-id option to the command:
$ openstack cluster profile list --full-id
+-------------------+----------+--------------------+---------------------+
| id | name | type | created_at |
+-------------------+----------+--------------------+---------------------+
| 560a8f9d-7596-... | myserver | os.nova.server-1.0 | 2015-05-05T13:26:00 |
| ceda64bd-70b7-... | mystack | os.heat.stack-1.0 | 2015-05-05T13:26:25 |
| 9b127538-a675-... | pstack | os.heat.stack-1.0 | 2015-06-25T12:59:01 |
+-------------------+----------+--------------------+---------------------+
The id column above contains the full UUID of profiles.
Sorting the List¶
You can specify the sorting keys and sorting direction when list profiles,
using the option --sort. The --sort option accepts a
string of format key1[:dir1],key2[:dir2],key3[:dir3], where the keys used
are profile properties and the dirs can be one of asc and desc. When
omitted, Senlin sorts a given key using asc as the default direction.
For example, the following command sorts the profiles using the name
property in descending order:
$ openstack cluster profile list --sort name:desc
When sorting the list of profiles, you can use one of type, name,
created_at and updated_at.
Paginating the List¶
In case you have a huge collection of profile objects, you can limit the
number of profiles returned from Senlin server, using the option
--limit. For example:
$ openstack cluster profile list --limit 1
+----------+----------+--------------------+---------------------+
| id | name | type | created_at |
+----------+----------+--------------------+---------------------+
| 560a8f9d | myserver | os.nova.server-1.0 | 2015-05-05T13:26:00 |
+----------+----------+--------------------+---------------------+
Yet another option you can specify is the ID of a profile object after which
you want to see the list starts. In other words, you don’t want to see those
profiles with IDs is or come before the one you specify. You can use the
option --marker for this purpose. For example:
$ openstack cluster profile list --limit 1 \
--marker ceda64bd-70b7-4711-9526-77d5d51241c5
+----------+--------+-------------------+---------------------+
| id | name | type | created_at |
+----------+--------+-------------------+---------------------+
| 9b127538 | pstack | os.heat.stack-1.0 | 2015-06-25T12:59:01 |
+----------+--------+-------------------+---------------------+
Creating a Profile¶
Before working with a Cluster or a Node, you will need a
Profile object created with a profile type. To create a profile, you
will need a “spec” file in YAML format. For example, below is a simple spec
for the os.heat.stack profile type (the source can be found in the
/examples/profiles/heat_stack/random_string/
heat_stack_random_string.yaml file).
type: os.heat.stack
version: 1.0
properties:
name: random_string_stack
template: random_string_stack.yaml
context:
region_name: RegionOne
The random_string_stack.yaml is the name of a Heat template file to be used
for stack creation. It is given here only as an example. You can
decide which properties to use based on your requirements.
Now you can create a profile using the following command:
$ cd /opt/stack/senlin/examples/profiles/heat_stack/random_string
$ openstack cluster profile create \
--spec heat_stack_random_string.yaml \
my_stack
+--------------+-----------------------------------------------------------+
| Property | Value |
+--------------+-----------------------------------------------------------+
| created_at | 2015-07-01T03:13:23 |
| id | c0389712-9c1a-4c58-8ba7-caa61b34b8b0 |
| metadata | {} |
| name | my_stack |
| spec | +------------+--------------------------------------------+ |
| | | property | value | |
| | +------------+--------------------------------------------+ |
| | | version | 1.0 | |
| | | type | "os.heat.stack" | |
| | | properties | { | |
| | | | "files": { | |
| | | | "file:///...": "<file contents>" | |
| | | | }, | |
| | | | "disable_rollback": true, | |
| | | | "template": { | |
| | | | "outputs": { | |
| | | | "result": { | |
| | | | "value": { | |
| | | | "get_attr": [ | |
| | | | "random", | |
| | | | "value" | |
| | | | ] | |
| | | | } | |
| | | | } | |
| | | | }, | |
| | | | "heat_template_version": "2014-10-16", | |
| | | | "resources": { | |
| | | | "random": { | |
| | | | "type": "OS::Heat::RandomString", | |
| | | | "properties": { | |
| | | | "length": 64 | |
| | | | } | |
| | | | } | |
| | | | }, | |
| | | | "parameters": { | |
| | | | "file": { | |
| | | | "default": { | |
| | | | "get_file": "file:///..." | |
| | | | }, | |
| | | | "type": "string" | |
| | | | } | |
| | | | } | |
| | | | }, | |
| | | | "parameters": {}, | |
| | | | "timeout": 60, | |
| | | | "environment": { | |
| | | | "resource_registry": { | |
| | | | "os.heat.server": "OS::Heat::Server" | |
| | | | } | |
| | | | }, | |
| | | | "context": { | |
| | | | "region_name": "RegionOne" | |
| | | | } | |
| | | | } | |
| | +------------+--------------------------------------------+ |
| type | os.heat.stack-1.0 |
+--------------+-----------------------------------------------------------+
From the outputs, you can see that the profile is created with a new id
generated. The spec property is dumped for the purpose of verification.
Optionally, you can attach some key-value pairs to the new profile when creating it. This data is referred to as the metadata for the profile:
$ openstack cluster profile create \
--spec heat_stack_random_string.yaml \
--metadata "author=Tom;version=1.0" \
my_stack
$ openstack cluster profile create \
--spec heat_stack_random_string.yaml \
--metadata author=Tom --metadata version=1.0 \
my_stack
Showing the Details of a Profile¶
Once there are profile objects in Senlin database, you can use the following command to show the properties of a profile:
$ openstack cluster profile show myserver
+------------+---------------------------------------------------------+
| Property | Value |
+------------+---------------------------------------------------------+
| created_at | 2015-07-01T03:18:58 |
| id | 70a36cc7-9fc7-460e-98f6-d44e3302e604 |
| metadata | {} |
| name | my_server |
| spec | +------------+----------------------------------------+ |
| | | property | value | |
| | +------------+----------------------------------------+ |
| | | version | 1.0 | |
| | | type | "os.nova.server" | |
| | | properties | { | |
| | | | "key_name": "oskey", | |
| | | | "flavor": 1, | |
| | | | "networks": [ | |
| | | | { | |
| | | | "network": "private" | |
| | | | } | |
| | | | ], | |
| | | | "image": "cirros-0.3.2-x86_64-uec", | |
| | | | "name": "cirros_server" | |
| | | | } | |
| | +------------+----------------------------------------+ |
| type | os.nova.server-1.0 |
+------------+---------------------------------------------------------+
Note that openstack cluster command line accepts one of the following values when retrieving a profile object:
- name: the name of a profile;
- ID: the UUID of a profile;
- short ID: an “abbreviated version” of the profile UUID.
Since Senlin doesn’t require a profile name to be unique, specifying profile
name for the openstack cluster profile show command won’t guarantee
that a profile object is returned. You may get a MultipleChoices exception
if more than one profile object match the name.
As another option, when retrieving a profile (or in fact any other objects, e.g. a cluster, a node, a policy etc.), you can specify the leading sub-string of an UUID as the “short ID” for query. For example:
$ openstack cluster profile show 560a8f9d
+----------+----------+--------------------+---------------------+
| id | name | type | created_at |
+----------+----------+--------------------+---------------------+
| 560a8f9d | myserver | os.nova.server-1.0 | 2015-05-05T13:26:00 |
+----------+----------+--------------------+---------------------+
$ openstack cluster profile show 560a
+----------+----------+--------------------+---------------------+
| id | name | type | created_at |
+----------+----------+--------------------+---------------------+
| 560a8f9d | myserver | os.nova.server-1.0 | 2015-05-05T13:26:00 |
+----------+----------+--------------------+---------------------+
As with query by name, a “short ID” won’t guarantee that a profile object is
returned even if it does exist. When there are more than one object matching
the short ID, you will get a MultipleChoices exception.
Updating a Profile¶
In general, a profile object should not be updated after creation. This is a
restriction to keep cluster and node status consistent at any time. However,
considering that there are cases where a user may want to change some
properties of a profile, openstack cluster command line does
support the profile update sub-command. For example, the following
command changes the name of a profile to new_server:
$ openstack cluster profile update --name new_server myserver
The following command creates or updates the metadata associated with the given profile:
$ openstack cluster profile update --metadata version=2.2 myserver
Changing the “spec” of a profile is not allowed. The only way to make a change is to create a new profile using the profile create sub-command.
Deleting a Profile¶
When there are no clusters or nodes referencing a profile object, you can delete it from the Senlin database using the following command:
$ openstack cluster profile delete myserver
Note that in this command you can use the name, the ID or the “short ID” to
specify the profile object you want to delete. If the specified criteria
cannot match any profiles, you will get a ResourceNotFound exception.If
more than one profile matches the criteria, you will get a MultipleChoices
exception. For example:
$ openstack cluster profile delete my
ERROR(404): The profile (my) could not be found.
Failed to delete any of the specified profile(s).
See Also¶
The following is a list of the links to documents related to profile’s creation and usage:
Except where otherwise noted, this document is licensed under Creative Commons Attribution 3.0 License. See all OpenStack Legal Documents.