Adds a new yaql function 'decryptData' which pairs with 'encryptData' on
the dashboard side.
Requires a valid secret storage backend (e.g. Barbican) to be configured
via Castellan in murano.conf, e.g.
auth_type = keystone_password
auth_url = <auth_url>
username = <username>
password = <password>
project_id = <project_id>
user_domain_name = <user_domain_name>
Murano will still work fine without this config but the encrypt/decrypt
functions will be unavailable.
Partially-Implements blueprint: allow-encrypting-of-muranopl-properties
Clients for various OpenStack services
were updated so that it would be
possible to specify region name explicitly.
If it is not provided then for clients that are
used internally by the core library it is taken
from the current region. Otherwise the default
environment region is used.
Magnum has changed its service name from 'containers' to 'container-infra'
which leads to failure of Magnum-app. So this fix changes the service
name in magnum_plugin.
To work with the latest Cloudify test environment, resource files
of NodeCellar example murano application should have version 3.2.1.
This patch fixes it in the app's README file, so user will use
correct version of files.
Magnum will remove all modules under magnumclient/openstack .
This plugin needs to update the import path.
Generated UIs of the non-muranoPL packages (HOT, cloudify_tosca and
CSAR) reference the type of object model's object by class FQNs.
However, when these packages are uploaded to glare, their class
contents are not indexed since the client which uploads them has no
knowledge of the generated classes.
To workaround this issue it is proposed to reference objects not just
by class FQNs but to include the package FQN and the version, so the
engine does not have to rely on Glare's indexes.
This patch fixes a wrong import statement in
contrib/plugins/magnum_plugin/magnum_plugin/__init__.py where it
tried to import client from magnumclient which is actually invalid.
HeatStack class expects to be owned by some other and uses it to find
OpenStack region name it belongs to. If there is no owner provided then
the default region is used which is random for multi-region setups.
MuranoPL code that was generated from HOT and TOSCA provided that owner
on "deploy" phase but not on "destroy". So the stack that was
successfully created earlier could never be deleted in multi-region
OpenStack installations for non-default region.
Currently Murano fails to delete Magnum bay due to missing
bay_ident attribute in case when Magnum bay status is CREATE_FAILED.
This patch fixs this issue.
This app uses python-magnumclient to deploy a Magnum Bay.
This bay can any of Kubernetes, Swarm and Mesos.
Partially-Implements: blueprint magnum-murano-rationalization
Add a custom plugin to support importing Murano application definition
archives of TOSCA CSAR files into Murano application catalog, and
deploy them through the catalog to Heat.
Co-Authored-By: Tetiana Lashchova <firstname.lastname@example.org>
Partially-Implements: blueprint support-tosca-format
The Environment class got two additional properties:
region - name of the region where environment should be deployed. If
null then home region is used.
regionConfigs - config property (taken from class config) that specifies
region parameters. It is a dictionary of a form regionName -> regionSettings.
Currently the only setting available is a agentRabbitMq that is itself
a dictionary containing the same options as [rabbitmq] section of murano.conf
Thus the configuration file io.murano.Environment.yaml may look like:
Implements Blueprint: assign-environment-to-region
* MuranoType and MuranoObjectParameterType smart types
were merged into a single smart type because their functionality overlap
to a large degree. New smart type is called MuranoObjectParameter
* Other smart types were renamed to have the same name pattern:
ThisParameterType -> ThisParameter,
InterfacesParameterType -> InterfacesParameter
* For MuranoObjectInterface instead of saying obj.data().propertyName
the syntax now is obj.properties.propertyName
* Single universal ClientManager class was dropped in favor of
of individual in-context methods to create OS clients without
* Environment class was renamed to ExecutionSession to avoid
common confusion with io.murano.Environment
* execution_session_local module was introduced to simplify
keep of per-execution session (per-deployment) data. This
is similar to thread-locals with the difference that there can
be many threads in single session.
* All OS-clients related code was migrated to keystone client
sessions and API v3 (except for GLARE and Mistral that doesn't
support sessions). This increases performance and solves
authentication problems that could be caused by token expiration
even with trusts enabled.
* [DEFAULT]/home_region setting was introduced instead of
[murano]/region_for_services to configure what region
should be used by the clients by default (where Murano API
resides). All client factories respect this setting.
Glare has been moved to a separate service and has been
renamed/refactored into glare namespace. This commit updates murano
Related bp: move-v3-to-glare