senlin/TODO

66 lines
1.7 KiB
Plaintext

High Priority
=============
DB
--
- Make sure cluster-policy association is deleted when a cluster is deleted
- Add field size to cluster table
- Modify node_set_status to check/update cluster status
- Add field cluster_id to event table so that query by cluster is easier
ENGINE
------
- cleanse scheduler module [Yanyan Hu]
- service start with rpc server(eventlet) and dispatcher(subthread);
- threadgroupmanager(TGM) module will now handle action;
- scheduler module will wrap TGM and expose interface to dispatcher,
e.g. run_action, wait, stop, notify;
- dispatcher using db.api to get action and use scheduler to dispatch
run and control action;
- use event to control action progress: wrap action_proc into ActionRunner
and using start/step to support event wait.
- action(ClusterAction, NodeAction, etc.) itself has to handle its own
progress control(e.g. implement as generator and support canceling)
- complete parser logic, construct profile/policy objects there?
DRIVER
------
- complete Heat stack driver [Qiming]
- add Heat resource driver
POLICY
------
- healthy policy[Liuh]
Middle Priority
===============
DB
--
- Add test cases for policy_delete with 'force' set to True[Liuh/ZhaiHF]
ENGINE
------
- Design and implement dynamical plugin loading mechanism that allows
loading plugins from any paths
OSLO
----
- Migrate to oslo.log
- Migrate to oslo.context
Low Priority
============
TEST
----
- Add test case in db cluster to test that cluster-policy association is
deleted when we delete a cluster
- Add test case to engine/parser
- Add test case to engine/registry
- Add test case to engine/environment