Peter Bacsko created YUNIKORN-1186:
--------------------------------------
Summary: Race condition during recovery
Key: YUNIKORN-1186
URL: https://issues.apache.org/jira/browse/YUNIKORN-1186
Project: Apache YuniKorn
Issue Type: Bug
Components: shim - kubernetes
Reporter: Peter Bacsko
Assignee: Peter Bacsko
The covery phase is not stable. Inside {{Context}}, we can create existing
{{Application}} objects in two possible ways:
1. Asynchronously from shared informers. When Yunikorn starts up, it
initializes the event handlers and it immediately receives the list of running
pods and populates the {{Context}} from {{Manager.addPod()}}.
2. Synchronously from {{Manager.ListApplications()}}.
We have to make sure that these two calls do not interfere with each other and
they create the same {{Application}} object, which right now does not seem to
be the case.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]