Sentinel & OpenSergo:下一代微服务治理标准与实践 赵奕豪（宿何） 阿里云 云原生技术专家，Sentinel & OpenSergo 负责人 Why 微服务治理？ 01 服务治理核心能力与原理揭秘 02 服务治理控制面与标准规范 03 未来展望 04 01 Why 微服务治理？ ’ alt=‘OCR图片’/> 现代微服务架构的挑战 用好微服务的挑战：稳定性 ’ alt=‘OCR图片’/> 针对指定的非致命错误异常自动重试，最大限度避免系统抖动 ’ alt=‘OCR图片’/> Sentinel：从流量防护到流量治理 Sentinel 诞生于阿里双十一流量峰值场景，逐步演进为基础技术模块，覆盖了所有的流量稳定性核心场景 Sentinel 正式开源，社区迅速发展，不断扩充生态与能力，逐步成为最受欢迎的同类组件 Sentinel 开始在多语言的生态中探索，推出 C++ 原生版本，同时针对 Service 全局流控的支持 Sentinel 推出 Go 原生版本，并不断与各个 Go 社区生态进行合作，继续朝着云原生方向演进 Sentinel Go: https://github.com/alibaba/sentinel- 云原生、智能化、标准化演进流量治理全面升级 ’ alt=‘OCR图片’/> Sentinel 2.0 流量治理大图 ’ alt=‘OCR图片’/> Sentinel Go 生态接入

高可用和集群架构与实践 主从复制-sentinel架构 高可用原理 发现原理基本切换原理 环境搭建 部署架构 网络规划 用户规划 持久化规划 目录规划 部署步骤 配置文件 维护操作 完整启动 启停redis 手动启动 启停sentinel 查看sentinel状态 查看master地址和端口 查看master配置 重置该sentinel 动态修改sentinel配置 主动切换 主从是否完全一致 接收所有事件信息 主实例宕测试 单从实例宕测试 双从实例宕测试 单sentinel宕测试 双sentinel宕测试 master所在主机整体宕测试 slave所在主机整体宕测试 脑裂测试 quorum测试 Master hang死测试 附：sentinel.conf被修改后的含义 附：sentinel事件含义 其他问题 只读性事件通知虚拟IP切换持久化动态修改 Sentinel最大连接数 Sharding架构（Redis3 使用alias方便操作 5.7.3 使用pdsh/pdcp进行多机器操作 5.7.4 使用脚本进行sentinel配置文件的备份 5.7 其他好用的配置技巧 5.7.1 使用supervisord进行进程管理 Supervisord是一个优秀的进程管理工具，一般在部署redis时采用它来进行redis、sentinel等进程的管理，一个已经在生产环境采用的supervisord配置文件如下： ; Sample

pairs_view> ## private: TBase vw; class inner_iterator { /* ... */ }; class inner sentinel { /* ... */ }; class inner_view { /* ... */ }; class outer_iterator { /* ... */ }; ## public: pairs_view> ## private: TBase vw; class inner iterator { /* ... */ }; class inner sentinel { /* ... */ }; class inner_view { /* ... */ }; class outer_iterator { /* ... */ }; ## public: pairs_view> ## private: TBase vw; class inner iterator { /* ... */ }; class inner sentinel { /* ... */ }; class inner_view { /* ... */ }; class outer_iterator { /* ... */ }; ## public:

jpg) ConfigMap  Sentinel  K8SAPI ![Im 微服务的治理阶段 • 修改sentinel.json 监控 • 框架Watch sentinel.json • 限流实时生效 • Sentinel监控 日志 链路 2021-08-22 16:36:03 INFO esentinel/component.go:76 read sentinel watch event {"comp": "core.sentinel", "compName": "sentinel", "event": "/Users/askuy/code/github/gotomicro/ego/examples/sentinel/http/config/sentinel.json", "path": "/Users/askuy/code/github/go

Sets the number of processor each worker can use. This does not affect auxiliary processes like the sentinel or monitor and is only useful for tweaking the performance of very high traffic clusters. The affinity process manager like Supervisor or Circus it is not strictly necessary. The cluster has an internal sentinel that checks the health of all the processes and recycles or reincarnates according to your settings Queue and checks the signing and unpacks the task. Before executing the task it sets a timer on the Sentinel indicating its about to start work. Afterwards it调整s time is reset and any results (including errors)

Queue and checks the signing and unpacks the task. Before executing the task it sets a timer on the Sentinel indicating its about to start work. Afterwards it调整s time is reset and any results (including errors) processed packages and saves both failed and successful packages to the Django database. ## Sentinel The sentinel spawns all process and then checks the health of all workers, including the pusher and the of a sudden death or timeout, it will reincarnate the failing processes. When a stop signal, the sentinel will halt the pusher and instruct the workers and monitor to finish the remaining items. See Stop

Sets the number of processor each worker can use. This does not affect auxiliary processes like the sentinel or monitor and is only useful for tweaking the performance of very high traffic clusters. The affinity process manager like Supervisor or Circus it is not strictly necessary. The cluster has an internal sentinel that checks the health of all the processes and recycles or reincarnates according to your settings Worker A worker process pulls a task of the Task Queue and it sets a shared countdown timer with Sentinel indicating it is about to start work. The worker then tries to execute the task and afterwards the

Sets the number of processor each worker can use. This does not affect auxiliary processes like the sentinel or monitor and is only useful for tweaking the performance of very high traffic clusters. The affinity [https://circus.readthedocs.org/en/latest/] it is not strictly necessary. The cluster has an internal sentinel that checks the health of all the processes and recycles or reincarnates according to your settings Queue and checks the signing and unpacks the task. Before executing the task it sets a timer on the Sentinel indicating its about to start work. Afterwards it the timer is reset and any results (including

Sets the number of processor each worker can use. This does not affect auxiliary processes like the sentinel or monitor and is only useful for tweaking the performance of very high traffic clusters. The affinity process manager like Supervisor or Circus it is not strictly necessary. The cluster has an internal sentinel that checks the health of all the processes and recycles or reincarnates according to your settings Worker A worker process pulls a task of the Task Queue and it sets a shared countdown timer with Sentinel indicating it is about to start work. The worker then tries to execute the task and afterwards the

Sets the number of processor each worker can use. This does not affect auxiliary processes like the sentinel or monitor and is only useful for tweaking the performance of very high traffic clusters. The affinity [https://circus.readthedocs.org/en/latest/] it is not strictly necessary. The cluster has an internal sentinel that checks the health of all the processes and recycles or reincarnates according to your settings Queue. A worker process pulls a task of the Task Queue and it sets a shared countdown timer with Sentinel indicating it is about to start work. The worker then tries to execute the task and afterwards the