b0636c1ed8
When a subcloud is first set to managed, the dcorch endpoint sync status keeps toggling between the two states - "in-sync" and "out-of-sync", before settling on "in-sync". This is because during the audit run, different endpoints (sysinv and identity) are done in separate greenthreads and each greenthread processes the resources (e.g. users, projects) in series. Thus, an enhancement is made to reduce the unnecessary messaging between dcorch and dcmanager, and to avoid the bouncing sync states. Change-Id: I98cec42eb5880d95d54182339dab6b5c951d6cfd Story: 2007267 Task: 38996 Signed-off-by: Jessica Castelino <jessica.castelino@windriver.com> |
||
---|---|---|
.. | ||
sync_services | ||
__init__.py | ||
dc_orch_lock.py | ||
fernet_key_manager.py | ||
generic_sync_manager.py | ||
initial_sync_manager.py | ||
quota_manager.py | ||
README.rst | ||
scheduler.py | ||
service.py | ||
subcloud.py | ||
sync_thread.py |
Service
Distributed Cloud Orchestration Engine Service has responsibility for:
Monitoring the subcloud status, and return the result if needed.
There is a single engine, with each subcloud having one persistent sync thread per endpoint-type. Sync audit threads will be created on demand as needed.
- service.py:
-
run orchengine service, and establish RPC server
- generic_sync_manager.py
-
Manages all the generic resource syncing.