Exchange_Server白手起家系列_系统SLA和监控流程

Click to edit Master title style,Click to edit Master text styles,Second level,Third level,Fourth level,Fifth level,*,*,Slide Title,Body Text,Second Level,Third Level,Fourth Level,Fifth Level,Exchange Server,白手起家系列系统,SLA,和监控流程,喻 勇,PMP/MCSE,微软最有价值专家，特约讲师,系列课程内容,万事开头难,进行企业,IT,架构评估,做一个,Exchange,管理员必须掌握的技术,-,软件篇,做一个,Exchange,管理员必须掌握的技术,-,硬件篇,邮件系统设计和选型的原则,如何执行压力测试？,软件安装和服务器管理的规范化,用户和系统迁移过程中的注意事项,大规模的用户帐号管理、组管理中采用的常见策略,客户端部署和管理,邮件系统日常维护的要点,系统,SLA,和监控流程,邮件安全、反病毒和反垃圾邮件,存储备份策略,你的系统可靠性有几个,9,？,Down,机了怎么办？,常见问题汇总和答疑,今日议程,SLA,是反映一项,IT,服务的综合指标，不仅仅代表系统的在线时间，更反映了用户对系统的认可程度。本次课程主要讨论构成,SLA,的组成部分，以及如何通过有效的工作来提升,SLA,和用户的满意程度。,SLA,的概念和深入解析,关键业务系统监控的流程,Exchange,主要的性能指标,性能优化建议,SLA,的概念和深入解析,SLA,：,Service Level Agreement,是在确定的成本下与客户达成的服务和系统使用情况的协议。,最简单的指标：系统服务在线时间,更进一步的衡量：,资源：软硬件的使用效率和投资回报,性能：用户的使用体验满意程度,服务：各类问题和帮助请求的解决速度,服务级别协议,(SLA),的结构,Service Based-One Service Level Agreement for a specific service for all customers of that service,Customer Based-One Service Level Agreement for a specific customer for all services,Multi-Level,Corporate Level-covering all generic Service Level Management(SLM)issues that would be appropriate for all customers,Customer Level-covering all Service Level Management(SLM)issues relevant to a specific customer group for all services used,Service Level-covering all Service Level Management(SLM)issues relevant to a specific service for a specific customer group,如何制定,SLA,在项目实施阶段的前期了解用户对,SLA,的需求,在实施过程中根据实际的软硬件投入情况进行动态的更新，并及时与客户,(,用户,),进行沟通,如果项目涉及到第三方供应商，应在制定,SLA,之前取得第三方的认可和确认,如何贯彻,SLA,服务器端,服务器配置和部署的规范统一,运营过程中遵循严格的流程,变更管理,最终用户,培训,Helpdesk,丰富客户端的功能，增强用户体验,整体协调,IT,支持部门内部的横向沟通和协调,管理与第三方供应商的关系,及时保持与客户,(IT,主管部门，,CIO,等,),的沟通,如何衡量,SLA,服务器在线时间,客户满意度,服务器在线时间,故障发生,检测到问题,检测时间,诊断,响应时间,修复,恢复服务,问题解决,下一次故障,修复时间,恢复时间,Mean Time to Repair(,MTTR,)“,总的停机时间,Time Between System Incidents,(,MTBSI,),Time Between Failures,(,MTBF,),客户满意度,客观的设定用户的期望值,建立良好的,Helpdesk,流程,如果是跨国企业，,Helpdesk,流程的一致性非常重要,性能,可扩展性,可用性,关键业务系统监控的流程,每日检查的内容,备份情况,邮件队列,系统性能,磁盘空间,硬件情况,昨天或昨晚的重要变更,每日检查报表的分发,建议使用,SharePoint,网站保存每日检查的报表,学会使用性能计数器来给,Exchange,把脉,demo,性能优化建议,Small design changes can reap big benefits,Do not be overly aggressive when tuning,especially when it involves registry changes,Always use the Microsoft Exchange Server Best Practices Analyzer Tool for validation,Storage Performance,The biggest cause of Exchange performance issues,Usually too few spindles to meet the demand,Exchange is I/O hungry,Back of the envelope:1 spindle for every 100 users,Actions to take,Work out your I/O profile,Tell your storage vendor how many IOPS you will produce,Set caching controllers to 100%write/4,kilobyte(,KB)cache page size,Think end-to-end storage(for example,Host Bus Adaptors),Use DiskPar to align tracks(performance boost),Use JetStress to validate your storage architecture,Read the paper,Optimizing Storage in Exchange 2003,!Warning Signs!,Average disk latency is greater than 20,milliseconds,Long disk latency spikes are greater than 50 milliseconds,Fast Access to Active Directory,If Exchange is slow accessing the Active Directory,directory service,store performance suffers and mail delivery is slow,Double check that the,Directory Service Access(,DSAccess)list only contains local domain controller or global catalog servers,Remember that round robin is used,Round-trip times should be less than 10 ms,Rule of thumb,You should have one global,catalog,processor for every four Exchange processors,Verify that the/3GB switch is being used on global,catalogs,Increases JET cache from 512 megabytes(MB)to 1 gigabyte(GB),More of the Active Directory is cached in memory,Correct Use of/3GB Switch,Mailbox and public folder servers,Implement when there is=1 GB RAM,Windows Server 2003 implement/USERVA=3030,Global catalog servers,Implement when there is=1 GB RAM,Front-end and routing servers,Do not implement regardless of memory,Instead,tune memory for system cache,/3GB and/USERVA Switches,User Mode,Kernel Mode,User Mode,User Mode,Kernel Mode,User Mode,HeapDeCommitFreeBlockThreshold,Should be set to 0 x40000 on all Exchange servers,Reduces virtual memory fragmentation by coalescing small free memory regions into larger blocks,Especially important for large Exchange mailbox servers 9582 be gone!,http:/ Hyper-Threading,Setting in BIOS,Sometimes defaults to OFF,Can improve Exchange processor efficiency by 25%,Works with Windows Server 2003 and Windows 2000 Server,Deploy Exchange Server 2003 SP2,30%reduction in I/O pattern,Drastic reduction in log file replay times,From 30 seconds to less than 2 seconds,Support for 8-way hyper-threading servers,Sets MaxDSNSize to 10 MB by default,Side:Detection of single-bit flip errors,Hardware Device Drivers,NonPaged Pool and page table entry(PTE)efficiency,Disable unused hardware devices,Especially consider network adapters and storage cards,Use identical hardware inside the server,One driver for multiple devices,Be especially careful with graphics adapters,Can result in serious PTE reduction,Keep drivers up-to-date,Switch to VGA.SYS driver,Need more PT