CloudDM 3.2.0 发布，统一登录入口和账号模型

{ "title_zh": "CloudDM 3.2：统一账号体系与 SQL 编辑器语言服务，团队数据库管理更省心", "body_zh": "# CloudDM 3.2：统一账号体系与 SQL 编辑器语言服务，团队数据库管理更省心\n\n数据库管理工具在团队协作场景下最让人头疼的两件事：一是权限和账号体系割裂，管理员和普通用户各走各的入口，维护成本高；二是 SQL 编辑器只能做语法高亮，补全和校验全靠人工。CloudDM 3.2.0 把这两个痛点一起解决了——统一登录入口和账号模型，并在 SQL 编辑器中引入后端插件驱动的语言服务。\n\n## 统一登录入口：不再区分管理员与普通用户\n\n3.2 之前，CloudDM 的管理员和普通用户走不同的登录流程，账号模型也是分开的。这在小团队里还能忍，一旦人数增长或角色频繁调整，每改一次权限就要在两套体系里同步，很容易出错。\n\n新版本把两者合并为单一账号模型：所有人从同一个入口登录，角色和权限通过统一的 RBAC 配置来区分。管理员只是拥有更多权限的账号，而不是另一种身份。这意味着——\n\n- 新成员入职只需创建一个账号，再分配角色即可，不再需要判断"该走哪个入口"。\n- 权限变更在一个面板里完成，减少遗漏。\n- 审计日志也统一了，谁在什么时间做了什么操作，一条链路可追溯。\n\n## SQL 编辑器语言服务：补全、分割、校验由后端插件驱动\n\n过去 CloudDM 的 SQL 编辑器主要做前端渲染，补全靠关键字字典，校验靠正则匹配。3.2 引入了数据源语言服务能力：后端插件根据具体数据源类型（MySQL、PostgreSQL、Redis 等）提供上下文感知的补全、语句分割和语法校验。\n\n实际效果：\n\n- 补全：输入 SELECT * FROM u 时，插件会根据当前数据库的 schema 列出以 u 开头的表和字段，而不是泛泛地提示关键字。\n- 分割：粘贴一段包含多条 SQL 的脚本，编辑器能按语句边界自动拆分，逐条执行或审核。\n- 校验：写完一条 DML 后，插件即时检查语法错误和潜在风险（比如缺少 WHERE 的 DELETE），在执行前就给出警告。\n\n这套语言服务是插件化的，不同数据源可以独立扩展，后续新增数据源类型时只需开发对应插件，不需要改动编辑器核心。\n\n## Docker 部署优化与快速上手\n\n3.2 对 Docker 部署流程做了精简，初始配置更少，启动更快。下面是一个最小化的部署示例，适合本地试用或小团队内部部署。\n\nyaml\n# docker-compose.yml — CloudDM 3.2 最小部署\nversion: '3.8'\nservices:\n clouddm:\n image: clouddm/clouddm:3.2.0\n container_name: clouddm\n ports:\n - \"8080:8080\"\n environment:\n # 统一账号模型后，只需配置管理员初始密码\n CLDDM_ADMIN_INIT_PASSWORD: \"ChangeMeOnProd!\"\n # 数据库连接（CloudDM 自身元数据存储）\n CLDDM_DB_HOST: \"clouddm-meta\"\n CLDDM_DB_PORT: \"5432\"\n CLDDM_DB_NAME: \"clouddm\"\n CLDDM_DB_USER: \"clouddm\"\n CLDDM_DB_PASSWORD: \"clouddm_meta_pass\"\n depends_on:\n - clouddm-meta\n restart: unless-stopped\n\n clouddm-meta:\n image: postgres:15-alpine\n container_name: clouddm-meta\n environment:\n POSTGRES_DB: clouddm\n POSTGRES_USER: clouddm\n POSTGRES_PASSWORD: clouddm_meta_pass\n volumes:\n - clouddm-meta-data:/var/lib/postgresql/data\n restart: unless-stopped\n\nvolumes:\n clouddm-meta-data:\n\n\n启动命令：\n\nbash\n# 拉取镜像并启动\ndocker compose up -d\n\n# 查看运行状态\ndocker compose ps\n\n# 访问 Web UI\n# 浏览器打开 http://localhost:8080\n# 用初始管理员密码登录，然后在统一面板中创建团队成员账号并分配角色\n\n\n部署后第一步操作建议：\n\n1. 用初始管理员密码登录，进入 团队管理 → 成员 创建普通用户账号。\n2. 在 角色权限 中为不同角色配置数据源访问范围和操作权限（只读、DML、DDL 等）。\n3. 添加数据源（MySQL / PostgreSQL / Redis 等），连接成功后打开 SQL 编辑器，输入表名前几个字母即可体验新的补全能力。\n\n## 采纳建议与注意事项\n\n- 迁移路径：如果你已经在用 CloudDM 3.1，升级到 3.2 后原有管理员账号会自动合并到统一模型中，但建议升级前备份元数据数据库，并在测试环境先跑一遍迁移脚本。\n- 权限重新梳理：统一模型意味着旧的"管理员入口专属权限"需要重新映射到 RBAC 角色上，升级后花半小时检查一遍角色配置，避免出现权限过大或过小的情况。\n- 语言服务插件：目前官方提供了 MySQL 和 PostgreSQL 的插件，其他数据源类型仍在开发中。如果你的团队重度使用 Redis 或 ClickHouse，可以关注后续版本或尝试自行开发插件（后端插件接口已开放）。\n- 生产部署：上面的 docker-compose 示例适合内网试用，生产环境请务必修改默认密码、启用 TLS、并将元数据数据库放在独立的高可用实例上。\n\nCloudDM 是开源免费的，代码在 GitHub 上可以直接获取。如果你的团队正在寻找一个带权限控制、数据脱敏和 SQL 审核能力的统一数据库管理平台，3.2 的统一账号和语言服务这两项改进值得实际跑一下。", "title_en": "CloudDM 3.2: Unified Account Model and SQL Language Services Make Team Database Management Easier", "body_en": "# CloudDM 3.2: Unified Account Model and SQL Language Services Make Team Database Management Easier\n\nTwo persistent headaches with database management tools in team settings: fragmented account models where admins and regular users go through separate login flows, and SQL editors that only do syntax highlighting with no real intelligence. CloudDM 3.2.0 tackles both — merging login flows into a single entry point with a unified account model, and introducing backend-plugin-driven language services for the SQL editor.\n\n## One Login, One Account Model\n\nBefore 3.2, CloudDM maintained separate login flows and account schemas for admins and regular users. That works in a five-person team, but as headcount grows and roles shift, every permission change requires syncing across two systems — a recipe for mistakes.\n\nThe new version collapses both into a single account model. Everyone logs in through the same entry point; roles and permissions are differentiated through unified RBAC configuration. An admin is simply an account with more privileges, not a different kind of identity. The practical upside:\n\n- Onboarding a new teammate means creating one account and assigning a role — no more deciding which login flow to use.\n- Permission changes happen in one panel, reducing gaps.\n- Audit trails are unified: who did what and when, one traceable chain.\n\n## SQL Editor Language Services: Completion, Splitting, Validation via Backend Plugins\n\nPreviously the SQL editor relied on frontend rendering — keyword dictionaries for completion, regex for validation. 3.2 adds data-source language service capabilities: backend plugins provide context-aware completion, statement splitting, and syntax validation tailored to the specific data source type (MySQL, PostgreSQL, Redis, etc.).\n\nWhat this looks like in practice:\n\n- Completion: typing SELECT * FROM u triggers the plugin to enumerate tables and columns starting with u based on the current database's actual schema, not a generic keyword list.\n- Splitting: paste a script with multiple SQL statements and the editor automatically identifies statement boundaries, enabling per-statement execution or review.\n- Validation: after writing a DML statement, the plugin immediately checks for syntax errors and risky patterns (e.g., a DELETE without a WHERE clause), surfacing warnings before execution.\n\nThe language service architecture is plugin-based. Different data sources get independent extensions; adding a new data source type later means developing a corresponding plugin without touching the editor core.\n\n## Docker Deployment Optimizations and Quick Start\n\n3.2 streamlines the Docker deployment flow — fewer initial config steps, faster startup. Below is a minimal deployment setup suitable for local trials or small-team internal deployments.\n\nyaml\n# docker-compose.yml — CloudDM 3.2 minimal deployment\nversion: '3.8'\nservices:\n clouddm:\n image: clouddm/clouddm:3.2.0\n container_name: clouddm\n ports:\n - \"8080:8080\"\n environment:\n # With the unified account model, just set the initial admin password\n CLDDM_ADMIN_INIT_PASSWORD: \"ChangeMeOnProd!\"\n # Database connection (CloudDM's own metadata store)\n CLDDM_DB_HOST: \"clouddm-meta\"\n CLDDM_DB_PORT: \"5432\"\n CLDDM_DB_NAME: \"clouddm\"\n CLDDM_DB_USER: \"clouddm\"\n CLDDM_DB_PASSWORD: \"clouddm_meta_pass\"\n depends_on:\n - clouddm-meta\n restart: unless-stopped\n\n clouddm-meta:\n image: postgres:15-alpine\n container_name: clouddm-meta\n environment:\n POSTGRES_DB: clouddm\n POSTGRES_USER: clouddm\n POSTGRES_PASSWORD: clouddm_meta_pass\n volumes:\n - clouddm-meta-data:/var/lib/postgresql/data\n restart: unless-stopped\n\nvolumes:\n clouddm-meta-data:\n\n\nStartup commands:\n\nbash\n# Pull images and start\ndocker compose up -d\n\n# Check status\ndocker compose ps\n\n# Access the Web UI\n# Open http://localhost:8080 in your browser\n# Log in with the initial admin password, then create team member accounts and assign roles from the unified panel\n\n\nRecommended first steps after deployment:\n\n1. Log in with the initial admin password, navigate to Team Management → Members, and create regular user accounts.\n2. In Role Permissions, configure data source access scopes and operation rights (read-only, DML, DDL, etc.) for each role.\n3. Add data sources (MySQL / PostgreSQL / Redis, etc.). Once connected, open the SQL editor and type the first few letters of a table name to experience the new completion capability.\n\n## Adoption Advice and Caveats\n\n- Migration path: If you're already running CloudDM 3.1, existing admin accounts automatically merge into the unified model on upgrade. Back up the metadata database beforehand and run the migration script in a test environment first.\n- Permission re-mapping: The unified model means old admin-exclusive permissions need to be re-mapped to RBAC roles. Spend half an hour after upgrade auditing role configurations to avoid over- or under-privileged accounts.\n- Language service plugins: Official plugins currently cover MySQL and PostgreSQL; other data source types are in development. If your team heavily uses Redis or ClickHouse, watch for future releases or try building your own plugin — the backend plugin interface is already open.\n- Production deployment: The docker-compose example above is for internal trials. For production, change default passwords, enable TLS, and place the metadata database on a dedicated high-availability instance.\n\nCloudDM is open-source and free, with code available on GitHub. If your team needs a unified database management platform with permission controls, data masking, and SQL review, the 3.2 improvements to accounts and language services make it worth a hands-on trial.", "seo_description_en": "CloudDM 3.2 unifies admin and user login into one account model and adds backend-plugin SQL language services for completion, splitting, and validation. Includes Docker quick-start guide." }