[SPARK-33716] Decommissioning Race Condition during Pod Snapshot

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.1.0, 3.2.0
• Fix Version/s: 3.2.0
• Component/s: Kubernetes
• Assignee: Holden Karau
• Reporter: Holden Karau
• Created: 08/Dec/20 21:05
• Updated: 09/Feb/21 18:23
• Resolved: 14/Dec/20 10:10

Description

Some version of Kubernetes may create a deletion timestamp field before changing the pod status to terminating, so a decommissioning node may have a deletion timestamp and a stage of running. Depending on when the K8s snapshot comes back this can cause a race condition with Spark believing the pod has been deleted before it has been.

Issue Links

• links to: [Github] Pull Request #30693 (holdenk)
• links to: [Github] Pull Request #30693 (holdenk)