[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-1-en-105":3,"doc-seo-239666-105":53,"doc-detail-239666-en":126},{"code":4,"msg":5,"data":6},0,"success",[7,14,19,24,29,34,39,44,49],{"id":8,"doc_module":9,"doc_module_name":10,"category_name":11,"show_sort_weight":12,"slug":13},11,1,"Template","Presentations",90,"presentations",{"id":15,"doc_module":9,"doc_module_name":10,"category_name":16,"show_sort_weight":17,"slug":18},12,"Resumes",80,"resumes",{"id":20,"doc_module":9,"doc_module_name":10,"category_name":21,"show_sort_weight":22,"slug":23},14,"Invoices",70,"invoices",{"id":25,"doc_module":9,"doc_module_name":10,"category_name":26,"show_sort_weight":27,"slug":28},15,"Posters",60,"posters",{"id":30,"doc_module":9,"doc_module_name":10,"category_name":31,"show_sort_weight":32,"slug":33},16,"Social Media",50,"social-media",{"id":35,"doc_module":9,"doc_module_name":10,"category_name":36,"show_sort_weight":37,"slug":38},17,"Forms",40,"forms",{"id":40,"doc_module":9,"doc_module_name":10,"category_name":41,"show_sort_weight":42,"slug":43},18,"Letters",30,"letters",{"id":45,"doc_module":9,"doc_module_name":10,"category_name":46,"show_sort_weight":47,"slug":48},21,"Paper Templates",5,"papers-templates",{"id":50,"doc_module":9,"doc_module_name":10,"category_name":51,"show_sort_weight":4,"slug":52},158,"General","general-158",{"code":4,"msg":54,"data":55},"ok",{"site_id":56,"language":57,"slug":58,"title":59,"keywords":60,"description":61,"schema_data":62,"social_meta":119,"head_meta":121,"extra_data":123,"updated_unix":125},105,"en","wg-batch-updates-whats-new-and-what-is-next","WG-Batch Updates: What’s New and What Is Next?","","WG-Batch updates outline the upcoming Kubernetes Job API and related controllers, highlighting work across versions 1.36–1.38. The content covers beta changes for suspended Jobs, including mutating Pod template resources and resetting scheduling startTime, plus alpha features to automatically create Workload and PodGroup resources. A major focus is JobSet’s Kubernetes-native grouping API, including Elastic JobSet and in-place restart, describing coordination, watermarks, barrier-based resync, and per-Pod agent behavior to preserve nodes, devices, sandboxes, and networking during worker failures.",{"@graph":63,"@context":118},[64,80,101],{"@type":65,"itemListElement":66},"BreadcrumbList",[67,71,74,77],{"item":68,"name":69,"@type":70,"position":9},"https://docshare.wps.com","Home","ListItem",{"item":72,"name":10,"@type":70,"position":73},"https://docshare.wps.com/template/",2,{"item":75,"name":11,"@type":70,"position":76},"https://docshare.wps.com/template/presentations/",3,{"item":78,"name":59,"@type":70,"position":79},"https://docshare.wps.com/template/wg-batch-updates-whats-new-and-what-is-next/239666/",4,{"url":78,"name":59,"@type":81,"image":82,"author":87,"headline":59,"publisher":90,"fileFormat":93,"inLanguage":57,"description":61,"dateModified":94,"datePublished":95,"encodingFormat":93,"isAccessibleForFree":96,"interactionStatistic":97},"DigitalDocument",{"url":83,"@type":84,"width":85,"height":86},"https://docshare.wps.com/thumbnails/wg-batch-updates-whats-new-and-what-is-next/239666.png","ImageObject",442,249,{"name":88,"@type":89},"Aurelia","Person",{"url":68,"name":91,"@type":92},"DocShare","Organization","application/pdf","2026-10-30","2026-09-11",true,{"@type":98,"interactionType":99,"userInteractionCount":15},"InteractionCounter",{"@type":100},"ViewAction",{"@type":102,"mainEntity":103},"FAQPage",[104,110,114],{"name":105,"@type":106,"acceptedAnswer":107},"What new Job API capabilities are planned or in progress?","Question",{"text":108,"@type":109},"Updates include beta support for mutating Pod template resources and resetting status.startTime for suspended Jobs. Alpha features also introduce Workload creation and PodGroup extensions via Job API scheduling details.","Answer",{"name":111,"@type":106,"acceptedAnswer":112},"What is JobSet and why is it important?",{"text":113,"@type":109},"JobSet is a Kubernetes-native API for managing groups of Jobs as a single unit, with a unified interface for deploying workload types such as HPC (MPI) and AI/ML training.",{"name":115,"@type":106,"acceptedAnswer":116},"How does In-Place Restart handle worker failures?",{"text":117,"@type":109},"On worker failure, JobSet resyncs and resumes without recreating all Pods. The workflow uses barriers to coordinate stale worker exit and restart containers in place, preserving node, devices, sandbox, and network contexts.","https://schema.org",{"og:url":78,"og:type":120,"og:title":59,"og:site_name":91,"og:description":61},"article",{"robots":122,"canonical":78},"index,follow",{"doc_id":124,"site_id":56},239666,1791724878,{"code":4,"msg":5,"data":127},{"doc_id":124,"user_id":128,"nickname":88,"user_avatar":129,"doc_module":9,"category_id":8,"category_name":11,"doc_title":59,"doc_description":61,"doc_content":130,"file_id":131,"file_url":132,"file_type":133,"file_size":134,"view_count":15,"is_deleted":4,"is_public":9,"is_downloadable":9,"audit_status":9,"page_count":135,"language":136,"language_code":57,"site_id":56,"html_lang":57,"table_of_contents":137,"faqs":138,"seo_title":139,"seo_description":61,"update_tm":140,"read_time":141},1099514068365,"https://ap-avatar.wpscdn.com/avatar/10000253d8d9f28188e?_k=1776742907772140068","WG-Batch Updates:  \nWhat’s New and What Is Next?  \nMichał Woźniak (@mimowo) Google  \nYuki Iwai (@tenzen-y) Preferred Networks, Inc.  \nJob   \nJob-Feature Graduations   \nMature Core K8SAPI, most features already in. Graduating:  \n1.36  \nMutablePodResourcesForSuspendedJobs—Beta  \nAllows changing Pod Template Spec resources for suspended Jobs  \nMutableSchedulingDirectivesForSuspendedJobs—Beta  \nAllows resetting status.startTime for suspended Jobs  \nWorkloadWithJob—Alpha1  \nAllows automatic creation of Workload and PodGroup resources for Job  \n1.37  \nWorkloadWithJob—Alpha2  \nExtend Job API with “.spec.scheduling” for WAS / TAS  \n1.38 (PLAN)  \nJobPodFailurePolicyName—Alpha  \nFine-grained JobPodFailurePolicy reason  \nJobSet  \nJobSet   \nJobSet is a Kubernetes-native API:  \n1. Managing a group of k8s Jobs as a unit.  \n2. A unified API to deploy the below workloads on Kubernetes  \na. HPC (e. g. , MPI) workloads  \nb. AI / ML training workloads (PyTorch, Jax, TensorFlow etc.)  \nv0 .12 Notable features:  \n● In-Place Restart  \n○ Cross-group effort:  \n■ sig-node × sig-apps × wg-batch  \n● Elastic JobSet  \n○ Mutable parallelisms and completions at the same timing while replicas is fixed.  \nNew Features-In-Place Restart  \n-motivation-  \nOne worker fails—the whole group resyncs  \nRecreate minutes at scale   \ndelete & reschedule every Pod  \nworker failure  \n→  \ndelete all Pods  \n→  \nreschedule + image pull, volume, device, network  \n→  \nall workers restart  \nInPlaceRestart  \ngoal: seconds  \nkeep Pods, restart containers  \nworker failure  \n→  \nrestart containers in place  \n→  \nNode, devices, sandbox & network preserved  \nworkers resync  \n→  \nNew Features-In-Place Restart  \n-flow over time-  \nFail, resync, resume  \nPhase 1 · startup  \n↓  \ncurrent = 0  \n▶ barriers lift, workers run  \nPhase 2 · Pod B fails  \n↓  \nprevious = 0  \nPod A is stale → exits  \nPhase 3 · resync  \n↓  \ncurrent = 1  \n▶ barriers lift, workers resume  \nNew Features-In-Place Restart  \n-components-  \nCoordinator and actuator  \nJobSet controller  \ncentral coordinator  \nList every Pod's attempt  \nCompute expected participant count Advance previous / current watermarks Fail the JobSet on maxRestarts  \nEvaluate FailurePolicy rules  \nin-place-restart-agent (NEW)  \nper-Pod actuator  \nWrite this Pod's attempt annotation Hold / lift the worker barrier  \nExit with a code to trigger restart Watch JobSet status for its generation Bypass when disabled  \nNew Features-In-Place Restart  \n-in-place-restart-agent deployment-  \nTwo agent modes  \nSidecar container  \nBarrier: agent startupProbe [HTTP](HTTP)  \nRestart: RestartAllContainers  \nMain container  \nBarrier: withhold worker command  \nRestart: Restart","cbCaiubNCxP4Pjtd","https://ap.wps.com/l/cbCaiubNCxP4Pjtd","pdf",2381023,53,"English","# WG-Batch Updates: What’s New and What Is Next?\n## Job API feature updates\n## JobSet overview and notable features\n## In-place restart motivation and flow\n## In-place restart components and agent deployment","[{\"question\":\"What new Job API capabilities are planned or in progress?\",\"answer\":\"Updates include beta support for mutating Pod template resources and resetting status.startTime for suspended Jobs. Alpha features also introduce Workload creation and PodGroup extensions via Job API scheduling details.\"},{\"question\":\"What is JobSet and why is it important?\",\"answer\":\"JobSet is a Kubernetes-native API for managing groups of Jobs as a single unit, with a unified interface for deploying workload types such as HPC (MPI) and AI/ML training.\"},{\"question\":\"How does In-Place Restart handle worker failures?\",\"answer\":\"On worker failure, JobSet resyncs and resumes without recreating all Pods. The workflow uses barriers to coordinate stale worker exit and restart containers in place, preserving node, devices, sandbox, and network contexts.\"}]","WG-Batch Updates: What’s New and What Is Next? | PDF",1789151708,19]