Anyone has implemented gang scheduling for Ray clusters with per ray cluster granularity. I do not need ”fancy” borrowing or fairness. But I do want per Ray cluster not per let’s say namespace shared quota. Any tips?

#Ray #Kueue #Volcano

Anyone using Kueue with Ray/Kuberay for gang scheduling?

#kuberay #kueue

#gb200 #gsoc #gsoc | Andrey Velichkevich We have exciting news just in time of KubeCon + CloudNativeCon – Kubeflow Trainer v2.1.0 is available! This release brings powerful new features, performance boosts, and enhanced flexibility for large...

Kubeflow Trainer v2.1.0 is released!
⚡️ Stream in-memory tabular data to GPUs from distributed cache with zero-copy transfer
🔥 Fine-tune LLMs on #Kubernetes with MLX on CUDA – now easier than ever!
🧠 Topology Aware Scheduling with #Kueue or #Volcano – essential for GB200
bit.ly/4qO8iM1

On démarre (enfin 😂) avec Laurent Grangeau qui nous présente #Kueue

GitOps et Jobs Scheduling | CNCF In-person Event - Édition d'avril 2025 du meetup Cloud Native Aix-Marseille.

🚨 Rendez-vous jeudi 24 avril sur Marseille pour une soirée 100% #CloudNative, avec du contenu technique pointu, des échanges passionnants et une ambiance conviviale !

🔥 Au programme : #Kueue, #GitOps, le tout sur #Kubernetes.

Merci à #Klanik pour l'accueil !

➡️ community.cncf.io/events/detai...

Pavneet with an awesome demo highlighting “near infinite” scale of batch job processing with #Kueue across 70K nodes on 80 clusters using #AKS #FleetManager, which by the way is based on the #KubeFleet project

How CERN analyzed 1 PetaByte per second using K8s with Ricardo Rocha One PetaByte is the equivalent of 11000 4k movies. And CERN's Large Hadron Collider (LHC) generates this every single second. Only a fraction of this data (~1 G

In our latest @pureperformance.bsky.social #podcast I had the chance to interview Ricardo Rocha from CERN. How they handle 1 PetaByte / Sec and how they use #kueue to distribute data analysis jobs across their global #k8s fleet.

🎧Tune in on Spreaker or your favorite podcast app: dt-url.net/k303wlp

Use ResourceFlavorReference instead of string consistently Use ResourceFlavorReference instead of string consistently

Want to contribute to Kubernetes? Fix inconsistent type usage by replacing string with ResourceFlavorReference consistently! Start by identifying scenarios where casting to string is still needed (e.g. NamespacedName). Help wanted!
#goodfirstissue #go #kubernetes-sigs #kueue

The second keynote of #KubeCon is a big one! Ricardo Rocha from Cern, and Marcin Wielgus from Google who works on Kueue, present "Multicluster Batch Jobs Dispatching with Kueue at CERN" #Kubernetes #Kueue

Ricardo Rocha from CERN and Marcin Wielgus from Google talking about the #Kueue project is used for multicluster batch job dispatching on #Kubernetes by facilitating unused resources are distributed and shared by workloads