<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Multi-Cluster on SREKubeCraft | Nick Nikolakakis</title><link>https://srekubecraft.io/tags/multi-cluster/</link><description>Recent content in Multi-Cluster on SREKubeCraft | Nick Nikolakakis</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Fri, 16 Oct 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://srekubecraft.io/tags/multi-cluster/index.xml" rel="self" type="application/rss+xml"/><item><title>Karmada v1.19 Multi-Component Scheduling - One Training Job, Three Clusters, Four Silent Failures</title><link>https://srekubecraft.io/posts/karmada-multi-component-scheduling/</link><pubDate>Fri, 16 Oct 2026 00:00:00 +0000</pubDate><guid>https://srekubecraft.io/posts/karmada-multi-component-scheduling/</guid><description>&lt;p&gt;You have three GPU clusters and one control plane in front of them. Someone submits a distributed training job: one master, two workers. You want the whole job in one cluster, because NCCL across regions is not a thing you do on purpose. You write the obvious policy: divide the replicas, pack them into as few clusters as possible.&lt;/p&gt;
&lt;p&gt;The job runs three times. Once in every cluster. Every status in the control plane is green.&lt;/p&gt;</description></item></channel></rss>