Kubernetes Scheduler AddGenericPodGroup Can Expose Partially Populated PodGroup
Is it fixed? Not yet.
Ongoing since Nov 17, 11:44 UTC. First reported Nov 17, 11:44 UTC.
What happened
When a PodGroup is observed after its pods, PriorityQueue.AddGenericPodGroup can make the root group poppable while it holds only a single member. This causes the scheduling cycle to run against a partial group because member order is random.
What caused it?
PriorityQueue.AddGenericPodGroup makes the root group poppable while it holds only one member when observed after its pods, and podGroupMemberPods.clear builds the pod list by ranging over a map leading to random member order.
No verified workaround yet
None of the linked reports describe a mitigation. This page updates when one appears.
Official fixTimeline
- Investigating(current)
- Workaround available(upcoming)
- Resolved(upcoming)
- First report on GitHub issue
scheduler: AddGenericPodGroup can expose a partially populated podGroup to Pop
Source
Evidence
1 report · 1 GitHubShow
- [1]GitHub First reportscheduler: AddGenericPodGroup can expose a partially populated podGroup to Pop(opens GitHub issue in a new tab)
What happened? When a PodGroup (or the last missing CompositePodGroup ancestor) is observed after its pods, PriorityQueue.AddGenericPodGroup can make the root group poppable while it holds only one member. The scheduling cycle then runs against a partial group. Step in pkg/scheduler/backend/queue: podGroupMemberPods.clear builds the pod list by ranging over a map, so member order is random. AddGenericPodGroup calls addPodGroupMember once per pod. For the first pod there is no queued root, so…
by helayoty 0upvotes 3comments
FAQ
What happened in the Kubernetes issue?
When a PodGroup is observed after its pods, PriorityQueue.AddGenericPodGroup can make the root group poppable while it holds only a single member. This causes the scheduling cycle to run against a partial group because member order is random.
What is affected by the Kubernetes issue?
Kubernetes.
When did the Kubernetes issue start and end?
It started Nov 17, 11:44 UTC and is still ongoing (1 d 16 h so far). This page updates as it develops.
How do I get alerts for Kubernetes?
Subscribe to the Kubernetes RSS feed for every new Kubernetes incident, or connect an AI assistant to WhatBrokeToday over MCP and ask whether Kubernetes is down. Kubernetes feed
Where does this information come from?
WhatBrokeToday collects public status pages, developer community reports and security advisories, groups the ones about the same problem, and links every report under Evidence on this page. This one has 1 report from 1 independent source.