Kubeservice博客

是非审之于己，毁誉听之于人，得失安之于数

TIPS之 Kubernetes GPU share 能力

Kubernetes GPU share 能力

Kubernetes GPU share 能力 GPU 软隔离模式通过 gpu-monitoring-tools 还获得 gpu device 驱动，并通过 deviceplugin 向kubelet注册GPU信息。底层通过 NVIDIA docker-smi 可对容器进行gpu分配 **GPU虚拟化技术:

Posted by 董江 on Tuesday, May 7, 2024

TIPS之 K8s docker exec 异常问题排查

K8s docker exec 异常问题排查

现象表现是：通过docker exec进入容器卡住，并且在10后 rpc timeout 报错背景信息 docker info 信息 [deployer@xxxx ~]$ sudo docker info Containers: 47 Running: 30 Paused: 0 Stopped: 17 Images: 30 Server Version: 18.09.5 Storage Driver: overlay2 Backing Filesystem: xfs Supports d_type: true

Posted by 董江 on Monday, July 31, 2023

TIPS之 Containerd 启动缓慢问题追查

Containerd 启动缓慢问题追查

一次Containerd 启动缓慢问题追查 div.notices { margin: 2rem 0; position: relative; } div.notices p { padding: 15px; white-space: pre-wrap; display: block; margin-top: 0rem; margin-bottom: 0rem; color: #666; } div.notices p:first-child:before { position: absolute; top: 2px; color: #fff; font-family: "Font Awesome 5 Free"; font-weight: 900; content: "\f06a"; left: 10px; } div.notices p:first-child:after { position: absolute;

Posted by Kubeservice博客 on Monday, July 3, 2023

TIPS之 Kubernetes kata-runtime 集群部署

Kubernetes kata-runtime 集群部署

Kubernetes kata-runtime 集群部署 Kata Containers 旨在构建一个安全且与 OCI 兼容的容器运行时，通过使用硬件虚拟化将每个容器工作负载放入轻量级虚拟机中，从而增强容器工作负载的安全性

Posted by 董江 on Wednesday, January 11, 2023

TIPS之 Kubernetes cgroup driver 变更方式

Kubernetes cgroup driver 变更方式

Kubernetes cgroup driver 变更方式现象 failed to create kubelet: misconfiguration: kubelet cgroup driver: "cgroupfs" is different from docker cgroup driver: "systemd" 文件驱动默认由systemd改成cgroupfs, 而我们安装的docker使用的文件驱动是

Posted by 董江 on Thursday, November 17, 2022

TIPS之 Kubernetes Node Containerd Runtime 问题排查

Kubernetes Node Containerd Runtime 排查

Kubernetes Node Containerd Runtime 问题排查 Containerd containerd 硬重启后，出现failed to recover state: failed to reserve sandbox name node1 containerd[80463]: time="2022-08-02T17:13:13.092422629Z" level=fatal msg="Failed to run CRI service" error="failed to recover state: failed to reserve sandbox name "kube-scheduler-node1_kube-system_705e7ce1217a37349a5567101e60165d_2": name "kube-scheduler-node1_kube-system_705e7ce1217a37349a5567101e60165d_2" is reserved for "139bb0ac7e050e9e28b994e78f651a8609f426f1b5bbfc887a0d4a3350b4eee2"" 日志很明显提升，容日中有一层

Posted by 董江 on Thursday, August 18, 2022