senlin/TODO.rst

HIGH PRIORITY

ENGINE

• cleanse scheduler module [Yanyan Hu]
• Node 'role' update may need to be propagated to the profile layer, because the profile may use it for changing the underlying physical object, e.g. the nova server that backs the node.

DRIVER

• Handle Heat stack operation exceptions [Qiming]

POLICY

• Formalize policy enforcement levels [Qiming]
• Enable placement policy and deletion policy to handle CLUSTER_RESIZE action.
• Investigate the impact of node-create and node-delete on certain policies.

TEST CASES

• Add test case the profile context can be saved and loaded correctly.

MIDDLE PRIORITY

API

• Revise the API for sorting, based on the following guideline: https://github.com/openstack/api-wg/blob/master/guidelines/pagination_filter_sort.rst
• According to the guidelines from API WG, we need to support page_reverse as a pagination parameter. https://review.openstack.org/190743
• Add support to replace a cluster node with another node
• Make object creation requests return code 202, since most creation are done asynchronously in Senlin.
• Make object creation requests return a location header set to the URI of the resource to be created. This is a requirement from API WG.
• API resource names should not include underscores. A guideline from API WG.
• Add API doc for CLUSTER_RESIZE operation.
• Add API doc for webhook APIs operation.
• Add support to have Senlin API run under Apache.

DB

• Add test cases for policy_delete with 'force' set to True[Liuh/ZhaiHF]
• The action data model is missing 'scheduled_start' and 'scheduled_stop' fields, we will need these fields for scheduled action execution.

ENGINE

• Add configuration option to enforce name uniqueness. There are reasonable requirements for cluster/node names to be unique within a project. This should be supported, maybe with the help from a name generator?
• Revise spec parser so that 'type' and 'version' are parts of the spec file This could be a client-only fix, or a client/server fix.
• Design and implement dynamical plugin loading mechanism that allows loading plugins from any paths
• Provide support to oslo.notification and allow nodes to receive and react to those notifications accordingly. [https://ask.openstack.org/en/question/46495/heat-autoscaling-adaptation-actions-on-existing-servers/]
• Allow actions to be paused and resumed. This is important for some background actions such as health checking
• Add support to template_url for heat stack profile Note: if template and template_url are both specified, use template Need to refer to heat api test for testing heat profile
• Revise start_action() in scheduler module so that it can handle cases when action_id specified is None. When action_id parameter is None, it means that the scheduler will pick a suitable READY action for execution.
• Add event logs wherever needed. Before that, we need a design on the criteria for events to be emitted.

OSLO

• Add support to oslo_versionedobjects
• Check if pre-context middleware needs logging and add special supports.

POLICY

• Scaling policy allowng a cluster to scale to existing nodes
• Healthy policy[Liuh]

DRIVER

• Add another abstract layer which hides interface differentiation between multiple drivers of the same type and provides unified interface for profile, e.g. alarm interfaces for scaling policy which can be mapped to both Ceilometer or Monasca driver; loadblancer interfaces for lb policy which can be mapped to both Neutron LBaaS or AWS LBaaS driver.

LOW PRIORITY

API

• Allow forced deletion of objects (cluster, node, policy, profile). The current problem is due to the limitations of the HTTP DELETE requests. We need to investigate whether a DELETE verb can carry query strings.

DRIVER

• add Heat resource driver
• add exception translation in driver

TEST

• Add test case to engine/parser
• Add test case to engine/registry
• Add test case to engine/environment

DOC

• Provide a sample conf file for customizing senlin options