《MicrosoftSQLServerDatabaseEngineIO》.docx

  1. 1、本文档共24页,可阅读全部内容。
  2. 2、原创力文档(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
  3. 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载
  4. 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
《MicrosoftSQLServerDatabaseEngineIO》.docx

Microsoft SQL Server Database Engine I/O by Bob Dorr, Microsoft SQL Server Principle Escalation Engineer, 1994 – Present Built: Jan 2008 Areas Covered Write Ahead Logging (WAL) Protocol Synchronous vs Asynchronous I/O Scatter / Gather I/O Sector alignment, Block Alignment Latching and a page: A read walk-through SQL Server I/O Sizes Data cache maintenance PAE and AWE Read Ahead User Mode and Kernel Mode (SYSTRAP) Sparse Files and Copy On Write (COW) Pages Locked Pages Scribbler(s) and Bit flips Page Protection and Constant Pages Checksum vs Torn Stale Read Stalled I/O WAL Protocol Write Ahead Logging ACID (Durability Property) Log records secured before data Hardened / Stable Media Log contains parity bit Commit Rollback Trigger Snapshot Synchronous vs Asynchronous I/O Sync: Wait for Completion Async: Post and Continue Overlapped Event Completion Port SQL Server 98% Async Usage Overlapped and HasOverlappedIoCompleted Network Layers Use Completion Port Backup/Restore Use Sync – Sequential Patterns dm_io_pending_io_requests Overlapped Structure Async Processing ~= CPU Package vs Phone Scatter / Gather I/O Consolidates or Distributes APIs ReadFileScatter WriteFileGather Increases Efficiency Used by SQL I/O Paths Used by Windows Page File Old Design: 6.x Sorting AWE Availability WriteMultiple # of 8K Pages Forward and Backward Buffer Pool Ramp-up Disk Memory Gather Scatter Sector Alignment Block Alignment Sector: Log Writes Block: Performance Avoid Crossovers DiskPart/DiskPar Utilities Discuss with your Vendor Double Touch Rewrites Defragment 4K Sectors Alignment: /kb/929491 To verify that an existing partition is aligned, divide the size of the stripe unit by the starting offset of the RAID disk group. Use the following syntax: ((Partition offset) * (Disk sector size)) / (Stripe unit size) Example of alignment calculations in bytes for a 256-KB stripe unit size: (63 * 512) / 262144 = 0.123046875 (64 * 512) / 262144 = 0.125 (128 * 512) /

文档评论(0)

jykt + 关注
实名认证
内容提供者

该用户很懒,什么也没介绍

1亿VIP精品文档

相关文档