搜索结果: "deployment"

tags: [kubernetes, architecture, deployment, pod, service, statefulset, networking, storage]

| **Controller Manager** | 运行各类控制器 | Deployment/ReplicaSet/Node 等控制器持续调谐 |

### Deployment — 无状态应用管理

kind: Deployment

**Deployment 控制器工作流程：**

tags: [kubernetes, deployment, troubleshooting, production, networking, security]

| 5 | 什么是 Deployment？ | 管理无状态应用，提供声明式更新、滚动回滚、自动修复 | — |

| 10 | ReplicaSet 与 Deployment？ | ReplicaSet 保证副本数，Deployment 管理 ReplicaSet 提供高级功能 | — |

| 11 | StatefulSet vs Deployment？ | StatefulSet 管理有状态应用，Pod 有固定身份；Deployment 管理无状态应用 | [[k8s-statefulset-guide]] |

| 60 | CI/CD 实现？ | Jenkins/GitLab CI/GitHub Actions + Argo CD（GitOps）| [[jenkins-multi-master-k8s-deployment]] |

tags: [kubernetes, statefulset, deployment, storage, networking, production]

## 与 Deployment 的本质区别

| 特性 | Deployment | StatefulSet |

StatefulSet 的滚动更新策略与 Deployment 不同：默认 OrderedReady（序号从大到小）。详见 [[k8s-rolling-update-pitfalls]]（滚动更新与非无损发布误区对比）。

以 MySQL 主从为例，对比 Deployment 和 StatefulSet：

- Created concepts: jenkins-multi-master-k8s-deployment

- Updated: k8s-statefulset-guide（StatefulSet vs Deployment 滚动更新对比）

- Updated: jenkins-multi-master-k8s-deployment, docker-image-optimization, k8s-rolling-update-pitfalls（添加 CI/CD 交叉引用）

| 2026-05-27 16:35 | create | jenkins-ansible-integration-guide | tags: devops,automation,case-study,deployment | source: 哇哈哈12138 2026-05-20 | P2 auto-publish, 3 cross-refs |

| 2026-05-27 17:58 | create | k8s-multicluster-istio-canary | tags: kubernetes,networking,deployment,architecture,monitoring,troubleshooting,automation,security | source: 小随小看 2026-05-15 | P2 auto-publish, 4 cross-refs |

> 🔙 **回滚：** 如果是新版镜像导致的，快速 `kubectl set image deployment/ =` 回退版本。

| `manifest unknown` | 镜像 tag 不存在 | `kubectl set image deployment/ =` |

| `authentication required` / `denied` | 私有仓库未认证 | `kubectl create secret docker-registry ...` → 关联到 Deployment |

> 🔙 **回滚：** `kubectl edit deployment ` 恢复旧镜像标签。

tags: [kubernetes, networking, deployment, architecture, monitoring, troubleshooting, automation, security]

1. **单集群灰度**：Deployment + Service + VS + DR 可用 + 自动回滚

- [[jenkins-multi-master-k8s-deployment]] — Jenkins 多 Master 架构

tags: [kubernetes, deployment, networking, troubleshooting, production, performance]

- 最终 Deployment 收敛到新版本

### 5.1 蓝绿发布（Blue-Green Deployment）

tags: [kubernetes, scheduling, production, deployment]

- 灰度 Deployment 通过 nodeAffinity 锁定灰度节点

- 验证通过后更新稳定版镜像 + 缩容灰度 Deployment

kubectl get deployment -n -o yaml > backup_.yaml

kubectl rollout history deployment -n

kubectl rollout undo deployment -n

tags: [kubernetes, troubleshooting, pod, deployment, production, networking, configmap]

# 或直接修改 Deployment 的 command 字段临时覆盖

- **认证失败:** 创建/更新 Secret → 关联到 Deployment 或 ServiceAccount

tags: [linux, docker, kubernetes, deployment, networking, security, monitoring, production]

| 5 | CI/CD 概念与环节？ | CI：频繁集成+自动化测试；CD：自动部署。环节：提交→构建→测试→质量检查→镜像构建→部署 | [[jenkins-multi-master-k8s-deployment]] |

tags: [devops, automation, case-study, deployment]

- [[jenkins-multi-master-k8s-deployment]] — Jenkins 多 Master 架构

tags: [kubernetes, deployment, architecture, build, networking, storage, monitoring, configmap, security, certificate, docker, container]

使用 StatefulSet 而非 Deployment，因为 Jenkins 需要**稳定的网络标识**和**独立的持久化存储**：

tags: [kubernetes, deployment, build, production]

- [[jenkins-multi-master-k8s-deployment]] — Jenkins 多 Master + K8S 部署方案

| **生效** | `kubectl rollout restart -n kube-system deployment/coredns` |

| **回滚** | `kubectl rollout undo -n kube-system deployment/coredns` |

kind: Deployment

| [[jenkins-multi-master-k8s-deployment]] | Jenkins 多 Master 部署中的 StatefulSet + PVC 实际案例 |

tags: [kubernetes, troubleshooting, security, hpa, deployment]

1. **增大 memory limits**（如果是正常业务增长）`kubectl patch deployment`

- deployment: 部署相关

tags: [docker, image, container, build, deployment, security]

kind: Deployment

tags: [kubernetes, database, deployment, production, architecture, mysql, statefulset]

业务接入层 → Deployment / StatefulSet / Job / Argo Rollout

tags: [kubernetes, troubleshooting, pod, scheduling, networking, storage, deployment]

**故障链：** 一个新 Deployment 因镜像 tag 写错，所有 Pod 反复 ImagePullBackOff → 同时大量拉取请求打满私有仓库带宽。

tags: [kubernetes, troubleshooting, pod, deployment, production, debugging]

# CoreDNS 以 Deployment 运行在 kube-system

kubectl get deployment -A

- CoreDNS Pod 异常：重启 `kubectl rollout restart deployment/coredns -n kube-system`

tags: [nginx, networking, production, deployment, ha, keepalived]

tags: [nginx, troubleshooting, networking, production, security, deployment]

tags: [nginx, deployment, production, security, networking]

配合 K8s Deployment：

- [[jenkins-multi-master-k8s-deployment]] — Jenkins on K8s（CI/CD pipeline 中 Terraform 的典型位置）

- [[jenkins-multi-master-k8s-deployment]] — Jenkins 多 Master 架构部署：K8S + Gateway API / StatefulSet / JCasC / 动态 Agent / VictoriaMetrics 监控