[YUNIKORN-1245] Clean up cache GetNodesInfoMap

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.1.0
• Component/s: shim - kubernetes
• Labels: pull-request-available
• Assignee: Craig Condit
• Reporter: Wilfred Spiegelenburg
• Created: 22/Jun/22 02:14
• Updated: 07/Feb/23 22:34
• Resolved: 28/Jun/22 03:27

Description

Clean up of the GetNodesInfoMap and GetNodesInfoMapCopy calls in the scheduler cache in the k8shim.

We should only have behavioural option and that is the locked one. The name can be chosen as either one. It must always return the clone of the map, not sure about the cloning the content at that point.

Not cloning each of the nodes at that point could give us a performance gain. All callers either filter or convert from a map to a slice. We can delay cloning the nodes itself to that point. This could save cloning a large number of objects as we never return them.

Issue Links

• links to: GitHub Pull Request #437