Persistent Memory in Kubernetes

on YouTube ↗22m

" summary

Persistent memory provides byte-addressable non-volatile memory that sits between DRAM and SSDs and enables much larger in-memory datasets with faster restarts, Intel® Optane™ Persistent Memory can scale to up to 6 TB per two-socket server. Administration uses ipmctl and ndctl; developers use PMDK and memkind; XFS/ext4 support direct PMEM mounts. Kubernetes integration is via the PMEM-CSI driver (raw block, ephemeral inline, operator, Docker Hub images), a scheduler extension and upcoming use of storage-capacity and generic-ephemeral APIs; core features are declared production-ready with Prometheus metrics.

AI-written summary · about

" chapters (6)

1. 0:00Intro
2. 0:46Persistent Memory (PMEM)
3. 3:06Using PMEM
4. 6:13Resources
5. 14:35Local Storage: PMEM-CSI
6. 16:46Local Storage: Kubernetes

" related talks

• 28mPersistent Storage with Kubernetes in Production - Which Solution and Why?Cheryl Hung · ContainerDays Hamburg 2018
• 01h 00mPMM on KubernetesDenys Kondratenko · Percona Community Live 2022
• 27mThe Storage Architecture of Intel's Data Management Platform (DMP)David Cohen, Phillip Reisner · Vault '19
• 23mKubernetes + Encrypted Memory = Security * PrivacyHarshal Patil, Pradipta Banerjee · KubeCon + CloudNativeCon Europe 2019
• 24mUsing Kubernetes with Data Processing Units to Offload InfrastructureThomas Phelan, Thomas Golway · KubeCon + CloudNativeCon North America 2021
• 19mPersistent Storage for Containers using GlusterRaghavendra Talur · DevConf 2017
• 28mKubernetes Storage Evolution: Enabling High Performance Distributed DatastoresErin A Boyd, Michelle Au · KubeCon + CloudNativeCon North America 2017
• 31mManaged Kubernetes — Next Gen Academic Infrastructure?Lukáš Hejtmánek, Viktória Spišaková · Kubernetes Batch + HPC Day North America 2022