{"id":37769395,"url":"https://github.com/yylt/etcdauto","last_synced_at":"2026-01-16T14:51:03.701Z","repository":{"id":332248326,"uuid":"1132587655","full_name":"yylt/etcdauto","owner":"yylt","description":null,"archived":false,"fork":false,"pushed_at":"2026-01-13T03:30:26.000Z","size":143,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-01-13T05:50:09.827Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Go","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/yylt.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":"CODEOWNERS","security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-01-12T07:08:47.000Z","updated_at":"2026-01-13T03:30:00.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/yylt/etcdauto","commit_stats":null,"previous_names":["yylt/etcdauto"],"tags_count":2,"template":false,"template_full_name":null,"purl":"pkg:github/yylt/etcdauto","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yylt%2Fetcdauto","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yylt%2Fetcdauto/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yylt%2Fetcdauto/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yylt%2Fetcdauto/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/yylt","download_url":"https://codeload.github.com/yylt/etcdauto/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yylt%2Fetcdauto/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28479402,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-16T11:59:17.896Z","status":"ssl_error","status_checked_at":"2026-01-16T11:55:55.838Z","response_time":107,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-01-16T14:51:02.945Z","updated_at":"2026-01-16T14:51:03.689Z","avatar_url":"https://github.com/yylt.png","language":"Go","funding_links":[],"categories":[],"sub_categories":[],"readme":"# etcd-cluster 自启动工具\n\n## 概述\n\n这是一个用于在 Kubernetes 环境中自动启动和管理 etcd 集群的工具。它能够根据当前集群状态自动决定是初始化新集群、加入现有集群，还是以学习者（learner）身份加入，并确保集群的高可用性和数据一致性。\n\n## 核心功能\n\n### 1. 集群状态检测\n- 通过 Kubernetes API 或 DNS 解析发现其他 etcd 节点\n- 使用健康检查（/readyz 端点）判断节点状态\n- 区分活跃节点和死亡节点\n\n### 2. 智能启动决策\n- **场景 1：无活跃节点** - 第一个节点（索引 0）初始化新集群\n- **场景 2：有活跃节点** - 其他节点加入现有集群\n- **场景 3：单主节点** - 以学习者身份加入然后升级为正式成员\n\n### 3. 故障恢复机制\n- 自动移除死亡节点\n- 重新加入数据丢失的成员\n- 学习者自动晋升\n\n## 环境变量要求\n\n### 必需环境变量\n| 变量名 | 说明 |\n|--------|------|\n| `SERVICE_NAME` | Kubernetes 服务名称 |\n| `MAX` | 集群最大节点数 |\n| `NODEIP_DIR` | 节点 IP 文件目录 |\n| `POD_NAME` | Pod 名称（格式：`prefix-index`） |\n| `POD_NAMESPACE` | Pod 所在命名空间 |\n| `POD_IPS` | Pod 的 IP 地址列表（逗号分隔） |\n| `ETCDCTL_CACERT` | etcd CA 证书路径 |\n| `ETCDCTL_CERT` | etcd 客户端证书路径 |\n| `ETCDCTL_KEY` | etcd 客户端密钥路径 |\n| `ETCD_DATA_DIR` | etcd 数据目录 |\n| `CLIENT_PORT` | etcd 客户端端口 |\n| `PEER_PORT` | etcd 对等端口 |\n\n### 可选环境变量\n| 变量名 | 说明 |\n|--------|------|\n| `LABELS` | 用于筛选 Pod 的标签选择器 |\n\n## 工作流程\n\n```mermaid\ngraph TD\n    A[开始] --\u003e B[检查环境变量]\n    B --\u003e C{环境变量是否完整?}\n    C --\u003e|否| D[退出程序]\n    C --\u003e|是| E[解析POD_NAME获取前缀和索引]\n    E --\u003e F[获取本机IP列表]\n    F --\u003e G[获取集群活跃节点]\n    \n    G --\u003e H{有活跃节点?}\n    H --\u003e|否| I{当前节点是索引0?}\n    I --\u003e|是| J[初始化新集群]\n    I --\u003e|否| K[等待重试]\n    \n    H --\u003e|是| L[获取etcd客户端]\n    L --\u003e M[检查当前成员状态]\n    \n    M --\u003e N{成员已存在?}\n    N --\u003e|是| O{数据目录存在?}\n    O --\u003e|是| P[直接启动etcd]\n    O --\u003e|否| Q[移除并重新加入]\n    \n    N --\u003e|否| R[检查集群规模]\n    \n    R --\u003e S{集群只有1个节点?}\n    S --\u003e|是| T[以学习者身份加入]\n    T --\u003e U[启动etcd后晋升]\n    S --\u003e|否| V[正常加入集群]\n    \n    V --\u003e W[启动etcd]\n    \n    J --\u003e X[启动etcd]\n    P --\u003e X\n    Q --\u003e V\n    U --\u003e X\n    W --\u003e X\n    \n    X --\u003e Y[等待退出信号]\n    Y --\u003e Z[程序退出]\n```\n\n## 详细流程说明\n\n### 1. 初始化阶段\n1. **环境验证**：检查所有必需环境变量\n2. **身份识别**：从 `POD_NAME` 解析节点前缀和索引号\n3. **IP 获取**：从 `POD_IPS` 获取本机 IP 列表\n\n### 2. 集群发现阶段\n1. **Kubernetes 发现**（优先）：使用 Kubernetes API 获取 Pod IP\n2. **DNS 回退**：如果 Kubernetes API 失败，通过 DNS 解析\n3. **IP 文件读取**：从 `NODEIP_DIR` 读取实际节点 IP\n4. **健康检查**：对每个节点进行 `/readyz` 端点检查\n\n### 3. 启动决策阶段\n\n#### 情况 A：无活跃节点\n- 只有索引为 0 的节点可以初始化新集群\n- 其他节点等待并重试\n\n#### 情况 B：有活跃节点\n1. **成员状态检查**：\n   - 如果成员已存在且数据完整：直接启动\n   - 如果成员已存在但数据丢失：移除后重新加入\n   - 如果成员不存在：作为新成员加入\n\n2. **加入策略**：\n   - **单节点集群**：以学习者身份加入，启动后晋升为正式成员\n   - **多节点集群**：正常加入\n   - **仲裁保护**：确保不会只剩一个非学习者节点\n\n### 4. etcd 启动阶段\n1. **环境变量设置**：\n   - `ETCD_INITIAL_CLUSTER_STATE`: \"new\" 或 \"existing\"\n   - `ETCD_INITIAL_CLUSTER`: 集群成员列表\n2. **数据目录清理**：确保干净的数据目录\n3. **进程启动**：启动 etcd 进程\n\n### 5. 监控与退出\n1. **信号处理**：捕获 SIGINT/SIGTERM 信号\n2. **进程监控**：监控 etcd 进程状态\n3. **优雅退出**：收到信号时终止 etcd 进程\n\n## 特殊处理逻辑\n\n### 学习者晋升\n当集群只有一个节点时，新节点先以学习者身份加入，启动 etcd 后立即晋升为正式成员，避免单点故障期间的仲裁丢失。\n\n### 死亡节点清理\n检测到在 Kubernetes/DNS 中不存在但仍在 etcd 集群中的成员时，自动将其移除。\n\n### 数据一致性保证\n如果成员存在于集群中但本地数据目录丢失，工具会先移除该成员，然后以新成员身份重新加入，确保数据一致性。\n\n## 日志输出\n\n工具使用结构化日志，关键信息包括：\n- 环境变量配置\n- 发现的活跃/死亡节点\n- 集群决策原因\n- etcd 启动参数\n- 进程状态变化\n\n## 部署建议\n\n1. **StatefulSet 部署**：确保稳定的网络标识和存储\n2. **资源限制**：为 etcd 配置足够的内存和 CPU\n3. **存储配置**：使用持久化存储确保数据安全\n4. **网络策略**：限制对 etcd 端点的访问\n5. **监控告警**：监控 etcd 集群健康状态\n\n## 故障排查\n\n### 常见问题\n1. **无法加入集群**：检查网络连通性和证书配置\n2. **仲裁丢失**：确保不会同时重启过多节点\n3. **数据不一致**：检查存储配置和备份策略\n\n### 日志关键词\n- \"Failed to get alive endpoints\"：集群发现失败\n- \"non learner is 1\"：仲裁保护触发\n- \"member is dead but in cluster\"：死亡节点清理\n- \"promote member success\"：学习者晋升成功\n\n## 安全性考虑\n\n1. **TLS 加密**：所有 etcd 通信使用 TLS\n2. **证书管理**：定期更新客户端证书\n3. **网络隔离**：限制对 etcd 端点的访问\n4. **权限控制**：使用最小权限原则配置服务账户\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fyylt%2Fetcdauto","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fyylt%2Fetcdauto","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fyylt%2Fetcdauto/lists"}