Red Hat Training

A Red Hat training course is available for Red Hat Enterprise Linux

群​​​​​​​集​​​​​​​套​​​​​​​件​​​​​​​概​​​​​​​述​​​​​​​

Red Hat Enterprise Linux 5

红​​​​​​​帽​​​​​​​企​​​​​​​业​​​​​​​版​​​​​​​ Linux 5 的​​​​​​​红​​​​​​​帽​​​​​​​群​​​​​​​集​​​​​​​套​​​​​​​件​​​​​​​

版 3

Logo

摘要

Red Hat Cluster Suite概​​​​​​​述​​​​​​​提​​​​​​​供​​​​​​​了​​​​​​​对​​​​​​​用​​​​​​​于​​​​​​​Red Hat Enterprise Linux 5 的​​​​​​​Red Hat Cluster Suite的​​​​​​​概​​​​​​​述​​​​​​​。​​​​​​​

简​​​​​​​介​​​​​​​

本​​​​​​​文​​​​​​​档​​​​​​​提​​​​​​​供​​​​​​​了​​​​​​​对​​​​​​​ Red Hat Enterprise Linux 5 中​​​​​​​ Red Hat Cluster Suite 的​​​​​​​高​​​​​​​级​​​​​​​概​​​​​​​述​​​​​​​,其​​​​​​​结​​​​​​​构​​​​​​​如​​​​​​​下​​​​​​​:
虽​​​​​​​然​​​​​​​本​​​​​​​文​​​​​​​档​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​只​​​​​​​是​​​​​​​一​​​​​​​个​​​​​​​概​​​​​​​述​​​​​​​,但​​​​​​​您​​​​​​​应​​​​​​​该​​​​​​​具​​​​​​​有​​​​​​​深​​​​​​​入​​​​​​​的​​​​​​​ Red Hat Enterprise Linux 的​​​​​​​相​​​​​​​关​​​​​​​知​​​​​​​识​​​​​​​并​​​​​​​理​​​​​​​解​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​计​​​​​​​算​​​​​​​的​​​​​​​概​​​​​​​念​​​​​​​,这​​​​​​​样​​​​​​​才​​​​​​​能​​​​​​​充​​​​​​​分​​​​​​​地​​​​​​​理​​​​​​​解​​​​​​​这​​​​​​​些​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
关​​​​​​​于​​​​​​​使​​​​​​​用​​​​​​​ Red Hat Enterprise Linux 的​​​​​​​更​​​​​​​多​​​​​​​信​​​​​​​息​​​​​​​,请​​​​​​​参​​​​​​​考​​​​​​​以​​​​​​​下​​​​​​​资​​​​​​​源​​​​​​​:
  • 《​​​​​​​Red Hat Enterprise Linux 安​​​​​​​装​​​​​​​指​​​​​​​南​​​​​​​》​​​​​​​ — 提​​​​​​​供​​​​​​​安​​​​​​​装​​​​​​​ Red Hat Enterprise Linux 5 的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
  • 《​​​​​​​Red Hat Enterprise Linux 部​​​​​​​署​​​​​​​指​​​​​​​南​​​​​​​》​​​​​​​ — 提​​​​​​​供​​​​​​​关​​​​​​​于​​​​​​​ Red Hat Enterprise Linux 5 的​​​​​​​部​​​​​​​署​​​​​​​、​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
有​​​​​​​关​​​​​​​ Red Hat Enterprise Linux 5 中​​​​​​​ Red Hat Cluster Suite 的​​​​​​​更​​​​​​​多​​​​​​​信​​​​​​​息​​​​​​​,请​​​​​​​参​​​​​​​考​​​​​​​以​​​​​​​下​​​​​​​资​​​​​​​源​​​​​​​:
  • 《​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​红​​​​​​​帽​​​​​​​群​​​​​​​集​​​​​​​系​​​​​​​统​​​​​​​》​​​​​​​ — 提​​​​​​​供​​​​​​​安​​​​​​​装​​​​​​​、​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​红​​​​​​​帽​​​​​​​群​​​​​​​集​​​​​​​组​​​​​​​件​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
  • Logical Volume Manager Administration — Provides a description of the Logical Volume Manager (LVM), including information on running LVM in a clustered environment.
  • 《​​​​​​​全​​​​​​​球​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​:配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​》​​​​​​​ — 提​​​​​​​供​​​​​​​安​​​​​​​装​​​​​​​、​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​维​​​​​​​护​​​​​​​红​​​​​​​帽​​​​​​​ GFS(红​​​​​​​帽​​​​​​​全​​​​​​​球​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​)的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
  • 《​​​​​​​全​​​​​​​球​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​ 2:配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​》​​​​​​​ — 提​​​​​​​供​​​​​​​安​​​​​​​装​​​​​​​、​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​维​​​​​​​护​​​​​​​红​​​​​​​帽​​​​​​​ GFS2(红​​​​​​​帽​​​​​​​全​​​​​​​球​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​ 2)的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
  • 《​​​​​​​使​​​​​​​用​​​​​​​设​​​​​​​备​​​​​​​映​​​​​​​射​​​​​​​器​​​​​​​多​​​​​​​路​​​​​​​径​​​​​​​》​​​​​​​ — 提​​​​​​​供​​​​​​​关​​​​​​​于​​​​​​​使​​​​​​​用​​​​​​​Red Hat Enterprise Linux 5 设​​​​​​​备​​​​​​​映​​​​​​​射​​​​​​​器​​​​​​​多​​​​​​​路​​​​​​​径​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
  • 《​​​​​​​在​​​​​​​全​​​​​​​球​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​里​​​​​​​使​​​​​​​用​​​​​​​ GNBD》​​​​​​​ — 个​​​​​​​关​​​​​​​于​​​​​​​在​​​​​​​红​​​​​​​帽​​​​​​​全​​​​​​​球​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​(GFS)里​​​​​​​使​​​​​​​用​​​​​​​全​​​​​​​球​​​​​​​网​​​​​​​路​​​​​​​块​​​​​​​设​​​​​​​备​​​​​​​(Global Network Block Device,GNBD)的​​​​​​​概​​​​​​​述​​​​​​​。​​​​​​​
  • 《​​​​​​​Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​管​​​​​​​理​​​​​​​》​​​​​​​ — 提​​​​​​​供​​​​​​​用​​​​​​​ Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​(Linux Virtual Server,LVS)配​​​​​​​置​​​​​​​高​​​​​​​性​​​​​​​能​​​​​​​的​​​​​​​系​​​​​​​统​​​​​​​和​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
  • 《​​​​​​​红​​​​​​​帽​​​​​​​群​​​​​​​集​​​​​​​套​​​​​​​件​​​​​​​发​​​​​​​行​​​​​​​注​​​​​​​记​​​​​​​》​​​​​​​ — 关​​​​​​​于​​​​​​​当​​​​​​​前​​​​​​​的​​​​​​​红​​​​​​​帽​​​​​​​群​​​​​​​集​​​​​​​套​​​​​​​件​​​​​​​(Red Hat Cluster Suite)的​​​​​​​发​​​​​​​行​​​​​​​注​​​​​​​记​​​​​​​。​​​​​​​
在​​​​​​​ Red Hat Enterprise Linux 文​​​​​​​档​​​​​​​ CD 和​​​​​​​在​​​​​​​线​​​​​​​网​​​​​​​站​​​​​​​ http://www.redhat.com/docs/ 里​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​找​​​​​​​到​​​​​​​ Red Hat Cluster Suite 文​​​​​​​档​​​​​​​和​​​​​​​其​​​​​​​他​​​​​​​ Red Hat 文​​​​​​​档​​​​​​​的​​​​​​​ HTML、​​​​​​​PDF 和​​​​​​​ PRM 版​​​​​​​本​​​​​​​。​​​​​​​

1. 反​​​​​​​馈​​​​​​​

如​​​​​​​果​​​​​​​您​​​​​​​发​​​​​​​现​​​​​​​了​​​​​​​排​​​​​​​印​​​​​​​错​​​​​​​误​​​​​​​或​​​​​​​者​​​​​​​有​​​​​​​改​​​​​​​进​​​​​​​本​​​​​​​文​​​​​​​档​​​​​​​的​​​​​​​建​​​​​​​议​​​​​​​,我​​​​​​​们​​​​​​​希​​​​​​​望​​​​​​​能​​​​​​​听​​​​​​​取​​​​​​​您​​​​​​​的​​​​​​​意​​​​​​​见​​​​​​​。​​​​​​​请​​​​​​​在​​​​​​​ Bugzilla(http://bugzilla.redhat.com/bugzilla/)里​​​​​​​提​​​​​​​交​​​​​​​关​​​​​​​于​​​​​​​组​​​​​​​件​​​​​​​ Documentation-cluster 的​​​​​​​报​​​​​​​告​​​​​​​。​​​​​​​
Be sure to mention the document's identifier:
Cluster_Suite_Overview(EN)-5 (2009-08-18T15:49)
By mentioning this document's identifier, we know exactly which version of the guide you have.
如​​​​​​​果​​​​​​​您​​​​​​​有​​​​​​​改​​​​​​​进​​​​​​​本​​​​​​​文​​​​​​​档​​​​​​​的​​​​​​​建​​​​​​​议​​​​​​​,请​​​​​​​尽​​​​​​​量​​​​​​​将​​​​​​​其​​​​​​​明​​​​​​​确​​​​​​​化​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​发​​​​​​​现​​​​​​​了​​​​​​​错​​​​​​​误​​​​​​​,请​​​​​​​指​​​​​​​出​​​​​​​章​​​​​​​节​​​​​​​号​​​​​​​以​​​​​​​及​​​​​​​其​​​​​​​周​​​​​​​围​​​​​​​的​​​​​​​相​​​​​​​关​​​​​​​文​​​​​​​字​​​​​​​,以​​​​​​​便​​​​​​​我​​​​​​​们​​​​​​​尽​​​​​​​快​​​​​​​找​​​​​​​到​​​​​​​并​​​​​​​更​​​​​​​正​​​​​​​该​​​​​​​错​​​​​​​误​​​​​​​。​​​​​​​

第 1 章 Red Hat Cluster Suite概​​​​​​​述​​​​​​​

群​​​​​​​集​​​​​​​系​​​​​​​统​​​​​​​为​​​​​​​关​​​​​​​键​​​​​​​产​​​​​​​品​​​​​​​服​​​​​​​务​​​​​​​提​​​​​​​供​​​​​​​可​​​​​​​靠​​​​​​​性​​​​​​​、​​​​​​​灵​​​​​​​活​​​​​​​性​​​​​​​和​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​。​​​​​​​使​​​​​​​用​​​​​​​ Red Hat Cluster Suite,您​​​​​​​可​​​​​​​以​​​​​​​创​​​​​​​建​​​​​​​群​​​​​​​集​​​​​​​来​​​​​​​满​​​​​​​足​​​​​​​对​​​​​​​性​​​​​​​能​​​​​​​、​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​、​​​​​​​负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​、​​​​​​​灵​​​​​​​活​​​​​​​性​​​​​​​、​​​​​​​文​​​​​​​件​​​​​​​共​​​​​​​享​​​​​​​以​​​​​​​及​​​​​​​经​​​​​​​济​​​​​​​性​​​​​​​的​​​​​​​需​​​​​​​要​​​​​​​。​​​​​​​本​​​​​​​章​​​​​​​提​​​​​​​供​​​​​​​ Red Hat Cluster Suite 组​​​​​​​件​​​​​​​和​​​​​​​功​​​​​​​能​​​​​​​的​​​​​​​概​​​​​​​述​​​​​​​,它​​​​​​​由​​​​​​​下​​​​​​​列​​​​​​​章​​​​​​​节​​​​​​​组​​​​​​​成​​​​​​​:

1.1. 群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​

群​​​​​​​集​​​​​​​是​​​​​​​由​​​​​​​两​​​​​​​个​​​​​​​或​​​​​​​多​​​​​​​个​​​​​​​计​​​​​​​算​​​​​​​机​​​​​​​(称​​​​​​​为​​​​​​​节​​​​​​​点​​​​​​​或​​​​​​​成​​​​​​​员​​​​​​​)组​​​​​​​成​​​​​​​的​​​​​​​一​​​​​​​起​​​​​​​执​​​​​​​行​​​​​​​任​​​​​​​务​​​​​​​的​​​​​​​系​​​​​​​统​​​​​​​。​​​​​​​主​​​​​​​要​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​类​​​​​​​型​​​​​​​有​​​​​​​ 4 种​​​​​​​:
  • 存​​​​​​​储​​​​​​​型​​​​​​​(Storage)
  • 高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​型​​​​​​​(High availability)
  • 负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​型​​​​​​​(Load balancing)
  • 高​​​​​​​性​​​​​​​能​​​​​​​型​​​​​​​(High performance)
存​​​​​​​储​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​跨​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​提​​​​​​​供​​​​​​​一​​​​​​​致​​​​​​​的​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​映​​​​​​​像​​​​​​​,它​​​​​​​允​​​​​​​许​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​同​​​​​​​步​​​​​​​地​​​​​​​读​​​​​​​取​​​​​​​和​​​​​​​写​​​​​​​入​​​​​​​单​​​​​​​个​​​​​​​的​​​​​​​共​​​​​​​享​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​。​​​​​​​存​​​​​​​储​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​将​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​的​​​​​​​安​​​​​​​装​​​​​​​和​​​​​​​修​​​​​​​补​​​​​​​限​​​​​​​制​​​​​​​到​​​​​​​一​​​​​​​个​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​,简​​​​​​​化​​​​​​​了​​​​​​​存​​​​​​​储​​​​​​​的​​​​​​​管​​​​​​​理​​​​​​​。​​​​​​​而​​​​​​​且​​​​​​​,采​​​​​​​用​​​​​​​群​​​​​​​集​​​​​​​范​​​​​​​围​​​​​​​的​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​,存​​​​​​​储​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​消​​​​​​​除​​​​​​​了​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​冗​​​​​​​余​​​​​​​拷​​​​​​​贝​​​​​​​并​​​​​​​简​​​​​​​化​​​​​​​了​​​​​​​备​​​​​​​份​​​​​​​和​​​​​​​恢​​​​​​​复​​​​​​​。​​​​​​​Red Hat Cluster Suite 是​​​​​​​通​​​​​​​过​​​​​​​ Red Hat GFS 提​​​​​​​供​​​​​​​存​​​​​​​储​​​​​​​群​​​​​​​集​​​​​​​功​​​​​​​能​​​​​​​的​​​​​​​。​​​​​​​
高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​通​​​​​​​过​​​​​​​消​​​​​​​除​​​​​​​单​​​​​​​点​​​​​​​失​​​​​​​效​​​​​​​以​​​​​​​及​​​​​​​在​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​停​​​​​​​止​​​​​​​运​​​​​​​作​​​​​​​时​​​​​​​将​​​​​​​服​​​​​​​务​​​​​​​从​​​​​​​一​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​切​​​​​​​换​​​​​​​到​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​,提​​​​​​​供​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​持​​​​​​​续​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​。​​​​​​​通​​​​​​​常​​​​​​​,在​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​中​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​读​​​​​​​写​​​​​​​数​​​​​​​据​​​​​​​(通​​​​​​​过​​​​​​​以​​​​​​​读​​​​​​​写​​​​​​​方​​​​​​​式​​​​​​​挂​​​​​​​载​​​​​​​的​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​)。​​​​​​​因​​​​​​​此​​​​​​​,高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​必​​​​​​​须​​​​​​​维​​​​​​​护​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​完​​​​​​​整​​​​​​​性​​​​​​​,因​​​​​​​为​​​​​​​一​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​是​​​​​​​从​​​​​​​另​​​​​​​一​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​接​​​​​​​管​​​​​​​某​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​控​​​​​​​制​​​​​​​。​​​​​​​客​​​​​​​户​​​​​​​端​​​​​​​中​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​中​​​​​​​产​​​​​​​生​​​​​​​故​​​​​​​障​​​​​​​的​​​​​​​节​​​​​​​点​​​​​​​在​​​​​​​群​​​​​​​集​​​​​​​外​​​​​​​并​​​​​​​不​​​​​​​可​​​​​​​见​​​​​​​。​​​​​​​(高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​有​​​​​​​时​​​​​​​候​​​​​​​被​​​​​​​称​​​​​​​作​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​群​​​​​​​集​​​​​​​。​​​​​​​)Red Hat Cluster Suite 通​​​​​​​过​​​​​​​其​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​(High-availability Service Management)组​​​​​​​件​​​​​​​提​​​​​​​供​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​功​​​​​​​能​​​​​​​。​​​​​​​
负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​将​​​​​​​网​​​​​​​络​​​​​​​服​​​​​​​务​​​​​​​请​​​​​​​求​​​​​​​分​​​​​​​摊​​​​​​​在​​​​​​​多​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​中​​​​​​​来​​​​​​​平​​​​​​​衡​​​​​​​请​​​​​​​求​​​​​​​负​​​​​​​载​​​​​​​。​​​​​​​因​​​​​​​为​​​​​​​您​​​​​​​可​​​​​​​以​​​​​​​按​​​​​​​照​​​​​​​负​​​​​​​载​​​​​​​来​​​​​​​调​​​​​​​整​​​​​​​节​​​​​​​点​​​​​​​数​​​​​​​目​​​​​​​,这​​​​​​​样​​​​​​​就​​​​​​​提​​​​​​​供​​​​​​​了​​​​​​​有​​​​​​​性​​​​​​​价​​​​​​​比​​​​​​​的​​​​​​​灵​​​​​​​活​​​​​​​性​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​中​​​​​​​的​​​​​​​某​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​不​​​​​​​可​​​​​​​操​​​​​​​作​​​​​​​了​​​​​​​,负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​软​​​​​​​件​​​​​​​检​​​​​​​测​​​​​​​到​​​​​​​这​​​​​​​个​​​​​​​故​​​​​​​障​​​​​​​并​​​​​​​将​​​​​​​请​​​​​​​求​​​​​​​重​​​​​​​定​​​​​​​向​​​​​​​到​​​​​​​其​​​​​​​他​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​中​​​​​​​的​​​​​​​节​​​​​​​点​​​​​​​故​​​​​​​障​​​​​​​在​​​​​​​群​​​​​​​集​​​​​​​外​​​​​​​是​​​​​​​不​​​​​​​可​​​​​​​见​​​​​​​的​​​​​​​。​​​​​​​Red Hat Cluster Suite 通​​​​​​​过​​​​​​​ Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​(Linux Virtual Server,LVS)提​​​​​​​供​​​​​​​了​​​​​​​负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​功​​​​​​​能​​​​​​​。​​​​​​​
高​​​​​​​性​​​​​​​能​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​使​​​​​​​用​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​来​​​​​​​执​​​​​​​行​​​​​​​并​​​​​​​行​​​​​​​运​​​​​​​算​​​​​​​。​​​​​​​高​​​​​​​性​​​​​​​能​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​允​​​​​​​许​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​以​​​​​​​并​​​​​​​行​​​​​​​的​​​​​​​方​​​​​​​式​​​​​​​工​​​​​​​作​​​​​​​,因​​​​​​​此​​​​​​​提​​​​​​​高​​​​​​​了​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​的​​​​​​​性​​​​​​​能​​​​​​​。​​​​​​​(高​​​​​​​性​​​​​​​能​​​​​​​型​​​​​​​群​​​​​​​集​​​​​​​也​​​​​​​被​​​​​​​称​​​​​​​为​​​​​​​计​​​​​​​算​​​​​​​群​​​​​​​集​​​​​​​或​​​​​​​网​​​​​​​格​​​​​​​计​​​​​​​算​​​​​​​。​​​​​​​)

注意

前​​​​​​​面​​​​​​​总​​​​​​​结​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​类​​​​​​​型​​​​​​​仅​​​​​​​反​​​​​​​映​​​​​​​了​​​​​​​基​​​​​​​本​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​;实​​​​​​​际​​​​​​​应​​​​​​​用​​​​​​​时​​​​​​​您​​​​​​​可​​​​​​​能​​​​​​​需​​​​​​​要​​​​​​​将​​​​​​​它​​​​​​​们​​​​​​​进​​​​​​​行​​​​​​​合​​​​​​​并​​​​​​​。​​​​​​​

1.2. Red Hat Cluster Suite Introduction

Red Hat Cluster Suite(RHCS)是​​​​​​​软​​​​​​​件​​​​​​​组​​​​​​​件​​​​​​​的​​​​​​​集​​​​​​​成​​​​​​​,它​​​​​​​可​​​​​​​以​​​​​​​不​​​​​​​同​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​进​​​​​​​行​​​​​​​部​​​​​​​署​​​​​​​来​​​​​​​满​​​​​​​足​​​​​​​您​​​​​​​对​​​​​​​性​​​​​​​能​​​​​​​、​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​、​​​​​​​负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​、​​​​​​​灵​​​​​​​活​​​​​​​性​​​​​​​、​​​​​​​文​​​​​​​件​​​​​​​共​​​​​​​享​​​​​​​和​​​​​​​经​​​​​​​济​​​​​​​性​​​​​​​的​​​​​​​需​​​​​​​要​​​​​​​。​​​​​​​
RHCS consists of the following major components (refer to 图 1.1 “Red Hat Cluster Suite Introduction”):
  • 群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​ — 提​​​​​​​供​​​​​​​节​​​​​​​点​​​​​​​以​​​​​​​群​​​​​​​集​​​​​​​方​​​​​​​式​​​​​​​一​​​​​​​起​​​​​​​工​​​​​​​作​​​​​​​所​​​​​​​需​​​​​​​的​​​​​​​基​​​​​​​本​​​​​​​功​​​​​​​能​​​​​​​、​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​管​​​​​​​理​​​​​​​、​​​​​​​成​​​​​​​员​​​​​​​资​​​​​​​格​​​​​​​管​​​​​​​理​​​​​​​ 、​​​​​​​锁​​​​​​​管​​​​​​​理​​​​​​​和​​​​​​​安​​​​​​​全​​​​​​​管​​​​​​​理​​​​​​​。​​​​​​​
  • 高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​ — 提​​​​​​​供​​​​​​​在​​​​​​​某​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​不​​​​​​​可​​​​​​​操​​​​​​​作​​​​​​​时​​​​​​​,服​​​​​​​务​​​​​​​从​​​​​​​一​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​到​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​。​​​​​​​
  • 群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​ — 用​​​​​​​于​​​​​​​设​​​​​​​置​​​​​​​、​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​Red Hat群​​​​​​​集​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​这​​​​​​​些​​​​​​​工​​​​​​​具​​​​​​​和​​​​​​​群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​组​​​​​​​件​​​​​​​、​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​和​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​组​​​​​​​件​​​​​​​、​​​​​​​存​​​​​​​储​​​​​​​组​​​​​​​件​​​​​​​一​​​​​​​起​​​​​​​使​​​​​​​用​​​​​​​。​​​​​​​
  • Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​(Linux Virtual Server,LVS)— 提​​​​​​​供​​​​​​​ IP 负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​的​​​​​​​路​​​​​​​由​​​​​​​软​​​​​​​件​​​​​​​。​​​​​​​LVS 在​​​​​​​一​​​​​​​对​​​​​​​冗​​​​​​​余​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​中​​​​​​​运​​​​​​​行​​​​​​​,它​​​​​​​将​​​​​​​客​​​​​​​户​​​​​​​请​​​​​​​求​​​​​​​均​​​​​​​匀​​​​​​​地​​​​​​​分​​​​​​​发​​​​​​​到​​​​​​​它​​​​​​​后​​​​​​​面​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​
您​​​​​​​可​​​​​​​为​​​​​​​Red Hat Cluster Suite提​​​​​​​供​​​​​​​以​​​​​​​下​​​​​​​组​​​​​​​件​​​​​​​,它​​​​​​​们​​​​​​​是​​​​​​​可​​​​​​​选​​​​​​​软​​​​​​​件​​​​​​​包​​​​​​​的​​​​​​​一​​​​​​​部​​​​​​​分​​​​​​​(但​​​​​​​不​​​​​​​是​​​​​​​Red Hat Cluster Suite的​​​​​​​一​​​​​​​部​​​​​​​分​​​​​​​):
  • Red Hat GFS(全​​​​​​​球​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​,Global File System)— 提​​​​​​​供​​​​​​​和​​​​​​​ Red Hat Cluster Suite 一​​​​​​​起​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​。​​​​​​​GFS 允​​​​​​​许​​​​​​​多​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​进​​​​​​​行​​​​​​​块​​​​​​​级​​​​​​​别​​​​​​​的​​​​​​​存​​​​​​​储​​​​​​​共​​​​​​​享​​​​​​​,就​​​​​​​好​​​​​​​像​​​​​​​每​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​都​​​​​​​连​​​​​​​接​​​​​​​至​​​​​​​本​​​​​​​地​​​​​​​存​​​​​​​储​​​​​​​空​​​​​​​间​​​​​​​一​​​​​​​样​​​​​​​。​​​​​​​
  • 群​​​​​​​集​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​(Cluster Logical Volume Manager,CLVM)— 提​​​​​​​供​​​​​​​群​​​​​​​集​​​​​​​存​​​​​​​储​​​​​​​的​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​。​​​​​​​

    注意

    When you create or modify a CLVM volume for a clustered environment, you must ensure that you are running the clvmd daemon. For further information, refer to 第 1.6 节 “群​​​​​​​集​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​”.
  • 全​​​​​​​球​​​​​​​网​​​​​​​络​​​​​​​块​​​​​​​设​​​​​​​备​​​​​​​(Global Network Block Device,GNBD)— GFS 的​​​​​​​辅​​​​​​​助​​​​​​​组​​​​​​​件​​​​​​​,它​​​​​​​将​​​​​​​块​​​​​​​级​​​​​​​别​​​​​​​的​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​导​​​​​​​出​​​​​​​至​​​​​​​以​​​​​​​太​​​​​​​网​​​​​​​络​​​​​​​。​​​​​​​这​​​​​​​是​​​​​​​在​​​​​​​ Red Hat GFS 中​​​​​​​使​​​​​​​用​​​​​​​块​​​​​​​级​​​​​​​别​​​​​​​存​​​​​​​储​​​​​​​的​​​​​​​一​​​​​​​个​​​​​​​经​​​​​​​济​​​​​​​的​​​​​​​方​​​​​​​法​​​​​​​。​​​​​​​
For a lower level summary of Red Hat Cluster Suite components and optional software, refer to 第 2 章 Red Hat Cluster Suite 组​​​​​​​件​​​​​​​摘​​​​​​​要​​​​​​​.
Red Hat Cluster Suite Introduction

图 1.1. Red Hat Cluster Suite Introduction

注意

图 1.1 “Red Hat Cluster Suite Introduction” includes GFS, CLVM, and GNBD, which are components that are part of an optional package and not part of Red Hat Cluster Suite.

1.3. Cluster Infrastructure

Red Hat Cluster Suite 群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​为​​​​​​​一​​​​​​​组​​​​​​​计​​​​​​​算​​​​​​​机​​​​​​​(称​​​​​​​为​​​​​​​节​​​​​​​点​​​​​​​或​​​​​​​成​​​​​​​员​​​​​​​)以​​​​​​​群​​​​​​​集​​​​​​​方​​​​​​​式​​​​​​​一​​​​​​​起​​​​​​​工​​​​​​​作​​​​​​​提​​​​​​​供​​​​​​​基​​​​​​​本​​​​​​​的​​​​​​​功​​​​​​​能​​​​​​​。​​​​​​​一​​​​​​​旦​​​​​​​使​​​​​​​用​​​​​​​群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​组​​​​​​​成​​​​​​​了​​​​​​​群​​​​​​​集​​​​​​​,您​​​​​​​就​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​其​​​​​​​他​​​​​​​Red Hat Cluster Suite的​​​​​​​组​​​​​​​件​​​​​​​来​​​​​​​满​​​​​​​足​​​​​​​群​​​​​​​集​​​​​​​需​​​​​​​求​​​​​​​(例​​​​​​​如​​​​​​​,设​​​​​​​立​​​​​​​在​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​上​​​​​​​进​​​​​​​行​​​​​​​文​​​​​​​件​​​​​​​共​​​​​​​享​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​,或​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​)。​​​​​​​群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​提​​​​​​​供​​​​​​​下​​​​​​​面​​​​​​​的​​​​​​​功​​​​​​​能​​​​​​​:
  • 群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​
  • 锁​​​​​​​管​​​​​​​理​​​​​​​
  • Fencing
  • 群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​管​​​​​​​理​​​​​​​

1.3.1. 群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​

Cluster management manages cluster quorum and cluster membership. CMAN (an abbreviation for cluster manager) performs cluster management in Red Hat Cluster Suite for Red Hat Enterprise Linux 5. CMAN is a distributed cluster manager and runs in each cluster node; cluster management is distributed across all nodes in the cluster (refer to 图 1.2 “CMAN/DLM Overview”).
CMAN keeps track of cluster quorum by monitoring the count of cluster nodes. If more than half the nodes are active, the cluster has quorum. If half the nodes (or fewer) are active, the cluster does not have quorum, and all cluster activity is stopped. Cluster quorum prevents the occurrence of a "split-brain" condition — a condition where two instances of the same cluster are running. A split-brain condition would allow each cluster instance to access cluster resources without knowledge of the other cluster instance, resulting in corrupted cluster integrity.
Quorum 由​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​间​​​​​​​经​​​​​​​由​​​​​​​以​​​​​​​太​​​​​​​网​​​​​​​的​​​​​​​消​​​​​​​息​​​​​​​通​​​​​​​信​​​​​​​决​​​​​​​定​​​​​​​的​​​​​​​。​​​​​​​Quorum 也​​​​​​​可​​​​​​​以​​​​​​​由​​​​​​​通​​​​​​​过​​​​​​​以​​​​​​​太​​​​​​​网​​​​​​​和​​​​​​​ quorum 磁​​​​​​​盘​​​​​​​的​​​​​​​消​​​​​​​息​​​​​​​的​​​​​​​组​​​​​​​合​​​​​​​来​​​​​​​决​​​​​​​定​​​​​​​。​​​​​​​对​​​​​​​于​​​​​​​经​​​​​​​由​​​​​​​以​​​​​​​太​​​​​​​网​​​​​​​的​​​​​​​ quorum,quorum 由​​​​​​​节​​​​​​​点​​​​​​​ vote 的​​​​​​​ 50% 加​​​​​​​ 1 组​​​​​​​成​​​​​​​的​​​​​​​。​​​​​​​而​​​​​​​对​​​​​​​于​​​​​​​经​​​​​​​由​​​​​​​ quorum 磁​​​​​​​盘​​​​​​​的​​​​​​​ quorum,quorum 由​​​​​​​用​​​​​​​户​​​​​​​指​​​​​​​定​​​​​​​的​​​​​​​条​​​​​​​件​​​​​​​组​​​​​​​成​​​​​​​。​​​​​​​

注意

在​​​​​​​默​​​​​​​认​​​​​​​情​​​​​​​况​​​​​​​下​​​​​​​,每​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​都​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​ quorum vote。​​​​​​​您​​​​​​​也​​​​​​​可​​​​​​​以​​​​​​​配​​​​​​​置​​​​​​​每​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​拥​​​​​​​有​​​​​​​多​​​​​​​个​​​​​​​ vote。​​​​​​​
CMAN 通​​​​​​​过​​​​​​​监​​​​​​​控​​​​​​​其​​​​​​​他​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​来​​​​​​​跟​​​​​​​踪​​​​​​​成​​​​​​​员​​​​​​​资​​​​​​​格​​​​​​​。​​​​​​​当​​​​​​​群​​​​​​​集​​​​​​​的​​​​​​​成​​​​​​​员​​​​​​​资​​​​​​​格​​​​​​​发​​​​​​​生​​​​​​​变​​​​​​​化​​​​​​​时​​​​​​​,群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​通​​​​​​​知​​​​​​​其​​​​​​​他​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​组​​​​​​​件​​​​​​​,它​​​​​​​们​​​​​​​再​​​​​​​采​​​​​​​取​​​​​​​合​​​​​​​适​​​​​​​的​​​​​​​行​​​​​​​动​​​​​​​。​​​​​​​例​​​​​​​如​​​​​​​,如​​​​​​​果​​​​​​​节​​​​​​​点​​​​​​​ A 加​​​​​​​入​​​​​​​了​​​​​​​群​​​​​​​集​​​​​​​并​​​​​​​挂​​​​​​​载​​​​​​​了​​​​​​​节​​​​​​​点​​​​​​​ B 和​​​​​​​ C已​​​​​​​经​​​​​​​挂​​​​​​​载​​​​​​​的​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​,那​​​​​​​么​​​​​​​节​​​​​​​点​​​​​​​ A 就​​​​​​​需​​​​​​​要​​​​​​​额​​​​​​​外​​​​​​​的​​​​​​​ journal 和​​​​​​​锁​​​​​​​的​​​​​​​管​​​​​​​理​​​​​​​来​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​在​​​​​​​规​​​​​​​定​​​​​​​的​​​​​​​时​​​​​​​间​​​​​​​内​​​​​​​没​​​​​​​有​​​​​​​传​​​​​​​送​​​​​​​消​​​​​​​息​​​​​​​,群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​将​​​​​​​将​​​​​​​这​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​从​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​删​​​​​​​除​​​​​​​,并​​​​​​​通​​​​​​​知​​​​​​​其​​​​​​​他​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​组​​​​​​​件​​​​​​​这​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​已​​​​​​​经​​​​​​​不​​​​​​​再​​​​​​​是​​​​​​​成​​​​​​​员​​​​​​​了​​​​​​​。​​​​​​​其​​​​​​​他​​​​​​​群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​组​​​​​​​件​​​​​​​在​​​​​​​获​​​​​​​得​​​​​​​这​​​​​​​个​​​​​​​通​​​​​​​知​​​​​​​后​​​​​​​将​​​​​​​决​​​​​​​定​​​​​​​采​​​​​​​取​​​​​​​什​​​​​​​么​​​​​​​样​​​​​​​的​​​​​​​行​​​​​​​动​​​​​​​。​​​​​​​例​​​​​​​如​​​​​​​,Fencing 将​​​​​​​保​​​​​​​护​​​​​​​不​​​​​​​再​​​​​​​是​​​​​​​成​​​​​​​员​​​​​​​的​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​
CMAN/DLM Overview

图 1.2. CMAN/DLM Overview

1.3.2. 锁​​​​​​​管​​​​​​​理​​​​​​​

Lock management is a common cluster-infrastructure service that provides a mechanism for other cluster infrastructure components to synchronize their access to shared resources. In a Red Hat cluster, DLM (Distributed Lock Manager) is the lock manager. As implied in its name, DLM is a distributed lock manager and runs in each cluster node; lock management is distributed across all nodes in the cluster (refer to 图 1.2 “CMAN/DLM Overview”). GFS and CLVM use locks from the lock manager. GFS uses locks from the lock manager to synchronize access to file system metadata (on shared storage). CLVM uses locks from the lock manager to synchronize updates to LVM volumes and volume groups (also on shared storage).

1.3.3. Fencing

Fencing is the disconnection of a node from the cluster's shared storage. Fencing cuts off I/O from shared storage, thus ensuring data integrity. The cluster infrastructure performs fencing through the fence daemon, fenced.
当​​​​​​​ CMAN 决​​​​​​​定​​​​​​​某​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​已​​​​​​​经​​​​​​​发​​​​​​​生​​​​​​​故​​​​​​​障​​​​​​​,它​​​​​​​将​​​​​​​告​​​​​​​知​​​​​​​其​​​​​​​他​​​​​​​群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​组​​​​​​​件​​​​​​​。​​​​​​​在​​​​​​​收​​​​​​​到​​​​​​​通​​​​​​​知​​​​​​​后​​​​​​​,fenced 将​​​​​​​保​​​​​​​护​​​​​​​(fence)故​​​​​​​障​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​其​​​​​​​他​​​​​​​群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​组​​​​​​​件​​​​​​​将​​​​​​​决​​​​​​​定​​​​​​​采​​​​​​​取​​​​​​​什​​​​​​​么​​​​​​​行​​​​​​​动​​​​​​​ — 也​​​​​​​就​​​​​​​是​​​​​​​说​​​​​​​,它​​​​​​​们​​​​​​​执​​​​​​​行​​​​​​​任​​​​​​​何​​​​​​​必​​​​​​​需​​​​​​​的​​​​​​​恢​​​​​​​复​​​​​​​。​​​​​​​例​​​​​​​如​​​​​​​,当​​​​​​​ DLM 和​​​​​​​ GFS 被​​​​​​​告​​​​​​​知​​​​​​​节​​​​​​​点​​​​​​​故​​​​​​​障​​​​​​​时​​​​​​​,将​​​​​​​暂​​​​​​​停​​​​​​​活​​​​​​​动​​​​​​​,直​​​​​​​到​​​​​​​它​​​​​​​们​​​​​​​检​​​​​​​测​​​​​​​到​​​​​​​ fenced 已​​​​​​​经​​​​​​​完​​​​​​​成​​​​​​​对​​​​​​​故​​​​​​​障​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​。​​​​​​​当​​​​​​​确​​​​​​​认​​​​​​​故​​​​​​​障​​​​​​​节​​​​​​​点​​​​​​​已​​​​​​​经​​​​​​​被​​​​​​​保​​​​​​​护​​​​​​​时​​​​​​​,DLM 和​​​​​​​ GFS 会​​​​​​​执​​​​​​​行​​​​​​​恢​​​​​​​复​​​​​​​。​​​​​​​DLM 释​​​​​​​放​​​​​​​对​​​​​​​失​​​​​​​败​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​锁​​​​​​​定​​​​​​​;GFS 恢​​​​​​​复​​​​​​​故​​​​​​​障​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​日​​​​​​​志​​​​​​​。​​​​​​​
保​​​​​​​护​​​​​​​程​​​​​​​序​​​​​​​根​​​​​​​据​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​决​​​​​​​定​​​​​​​使​​​​​​​用​​​​​​​哪​​​​​​​种​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​。​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​用​​​​​​​两​​​​​​​个​​​​​​​关​​​​​​​键​​​​​​​元​​​​​​​素​​​​​​​定​​​​​​​义​​​​​​​了​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​:保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​(fencing agent)和​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​(fencing device)。​​​​​​​保​​​​​​​护​​​​​​​程​​​​​​​序​​​​​​​调​​​​​​​用​​​​​​​这​​​​​​​里​​​​​​​定​​​​​​​义​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​而​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​则​​​​​​​通​​​​​​​过​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​来​​​​​​​保​​​​​​​护​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​当​​​​​​​保​​​​​​​护​​​​​​​过​​​​​​​程​​​​​​​结​​​​​​​束​​​​​​​时​​​​​​​,保​​​​​​​护​​​​​​​程​​​​​​​序​​​​​​​将​​​​​​​通​​​​​​​知​​​​​​​群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​。​​​​​​​
Red Hat Cluster Suite提​​​​​​​供​​​​​​​各​​​​​​​种​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​:
  • 电​​​​​​​源​​​​​​​保​​​​​​​护​​​​​​​(Power fencing)— 该​​​​​​​方​​​​​​​法​​​​​​​使​​​​​​​用​​​​​​​电​​​​​​​源​​​​​​​控​​​​​​​制​​​​​​​器​​​​​​​来​​​​​​​关​​​​​​​闭​​​​​​​不​​​​​​​可​​​​​​​操​​​​​​​作​​​​​​​的​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​
  • 光​​​​​​​线​​​​​​​通​​​​​​​道​​​​​​​开​​​​​​​关​​​​​​​保​​​​​​​护​​​​​​​(Fibre Channel switch fencing)— 该​​​​​​​方​​​​​​​法​​​​​​​禁​​​​​​​用​​​​​​​连​​​​​​​接​​​​​​​至​​​​​​​不​​​​​​​可​​​​​​​操​​​​​​​作​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​光​​​​​​​线​​​​​​​通​​​​​​​道​​​​​​​端​​​​​​​口​​​​​​​。​​​​​​​
  • GNBD fencing — A fencing method that disables an inoperable node's access to a GNBD server.
  • 其​​​​​​​他​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​ — 如​​​​​​​禁​​​​​​​用​​​​​​​ I/O 或​​​​​​​不​​​​​​​可​​​​​​​操​​​​​​​作​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​电​​​​​​​源​​​​​​​,包​​​​​​​括​​​​​​​ IBM Bladecenters, PAP, DRAC/MC, HP ILO, IPMI, IBM RSA II 等​​​​​​​。​​​​​​​
图 1.3 “Power Fencing Example” shows an example of power fencing. In the example, the fencing program in node A causes the power controller to power off node D. 图 1.4 “Fibre Channel Switch Fencing Example” shows an example of Fibre Channel switch fencing. In the example, the fencing program in node A causes the Fibre Channel switch to disable the port for node D, disconnecting node D from storage.
Power Fencing Example

图 1.3. Power Fencing Example

Fibre Channel Switch Fencing Example

图 1.4. Fibre Channel Switch Fencing Example

指​​​​​​​定​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​通​​​​​​​过​​​​​​​编​​​​​​​辑​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​完​​​​​​​成​​​​​​​,如​​​​​​​分​​​​​​​配​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​名​​​​​​​、​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​(fencing agent)、​​​​​​​以​​​​​​​及​​​​​​​用​​​​​​​于​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​每​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​(fencing device)。​​​​​​​
The way in which a fencing method is specified depends on if a node has either dual power supplies or multiple paths to storage. If a node has dual power supplies, then the fencing method for the node must specify at least two fencing devices — one fencing device for each power supply (refer to 图 1.5 “Fencing a Node with Dual Power Supplies”). Similarly, if a node has multiple paths to Fibre Channel storage, then the fencing method for the node must specify one fencing device for each path to Fibre Channel storage. For example, if a node has two paths to Fibre Channel storage, the fencing method should specify two fencing devices — one for each path to Fibre Channel storage (refer to 图 1.6 “Fencing a Node with Dual Fibre Channel Connections”).
Fencing a Node with Dual Power Supplies

图 1.5. Fencing a Node with Dual Power Supplies

Fencing a Node with Dual Fibre Channel Connections

图 1.6. Fencing a Node with Dual Fibre Channel Connections

您​​​​​​​可​​​​​​​以​​​​​​​配​​​​​​​置​​​​​​​节​​​​​​​点​​​​​​​具​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​或​​​​​​​多​​​​​​​个​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​。​​​​​​​当​​​​​​​您​​​​​​​配​​​​​​​置​​​​​​​某​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​具​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​时​​​​​​​,这​​​​​​​个​​​​​​​方​​​​​​​法​​​​​​​就​​​​​​​是​​​​​​​保​​​​​​​护​​​​​​​该​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​唯​​​​​​​一​​​​​​​可​​​​​​​用​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​。​​​​​​​而​​​​​​​当​​​​​​​您​​​​​​​配​​​​​​​置​​​​​​​某​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​具​​​​​​​有​​​​​​​多​​​​​​​个​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​时​​​​​​​,保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​将​​​​​​​按​​​​​​​照​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​里​​​​​​​指​​​​​​​定​​​​​​​的​​​​​​​顺​​​​​​​序​​​​​​​挨​​​​​​​个​​​​​​​进​​​​​​​行​​​​​​​层​​​​​​​叠​​​​​​​(cascade)。​​​​​​​如​​​​​​​果​​​​​​​某​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​出​​​​​​​现​​​​​​​故​​​​​​​障​​​​​​​,配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​里​​​​​​​指​​​​​​​定​​​​​​​的​​​​​​​第​​​​​​​一​​​​​​​个​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​将​​​​​​​对​​​​​​​它​​​​​​​进​​​​​​​行​​​​​​​保​​​​​​​护​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​第​​​​​​​一​​​​​​​个​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​没​​​​​​​有​​​​​​​成​​​​​​​功​​​​​​​,指​​​​​​​定​​​​​​​的​​​​​​​第​​​​​​​二​​​​​​​个​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​将​​​​​​​被​​​​​​​使​​​​​​​用​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​所​​​​​​​有​​​​​​​方​​​​​​​法​​​​​​​都​​​​​​​没​​​​​​​有​​​​​​​成​​​​​​​功​​​​​​​,保​​​​​​​护​​​​​​​将​​​​​​​从​​​​​​​第​​​​​​​一​​​​​​​个​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​重​​​​​​​新​​​​​​​开​​​​​​​始​​​​​​​,且​​​​​​​安​​​​​​​装​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​里​​​​​​​指​​​​​​​定​​​​​​​的​​​​​​​顺​​​​​​​序​​​​​​​进​​​​​​​行​​​​​​​循​​​​​​​环​​​​​​​,直​​​​​​​到​​​​​​​这​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​被​​​​​​​成​​​​​​​功​​​​​​​地​​​​​​​保​​​​​​​护​​​​​​​为​​​​​​​止​​​​​​​。​​​​​​​

1.3.4. 群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​系​​​​​​​统​​​​​​​

The Cluster Configuration System (CCS) manages the cluster configuration and provides configuration information to other cluster components in a Red Hat cluster. CCS runs in each cluster node and makes sure that the cluster configuration file in each cluster node is up to date. For example, if a cluster system administrator updates the configuration file in Node A, CCS propagates the update from Node A to the other nodes in the cluster (refer to 图 1.7 “CCS Overview”).
CCS Overview

图 1.7. CCS Overview

Other cluster components (for example, CMAN) access configuration information from the configuration file through CCS (refer to 图 1.7 “CCS Overview”).
Accessing Configuration Information

图 1.8. Accessing Configuration Information

群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​(/etc/cluster/cluster.conf)是​​​​​​​一​​​​​​​个​​​​​​​ XML 文​​​​​​​件​​​​​​​,它​​​​​​​描​​​​​​​述​​​​​​​了​​​​​​​下​​​​​​​列​​​​​​​群​​​​​​​集​​​​​​​属​​​​​​​性​​​​​​​:
  • 群​​​​​​​集​​​​​​​名​​​​​​​ — 显​​​​​​​示​​​​​​​群​​​​​​​集​​​​​​​的​​​​​​​名​​​​​​​字​​​​​​​、​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​版​​​​​​​本​​​​​​​、​​​​​​​当​​​​​​​节​​​​​​​点​​​​​​​加​​​​​​​入​​​​​​​群​​​​​​​集​​​​​​​或​​​​​​​被​​​​​​​保​​​​​​​护​​​​​​​所​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​基​​​​​​​本​​​​​​​保​​​​​​​护​​​​​​​时​​​​​​​间​​​​​​​属​​​​​​​性​​​​​​​。​​​​​​​
  • 群​​​​​​​集​​​​​​​ — 显​​​​​​​示​​​​​​​群​​​​​​​集​​​​​​​的​​​​​​​每​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​,指​​​​​​​定​​​​​​​节​​​​​​​点​​​​​​​名​​​​​​​、​​​​​​​节​​​​​​​点​​​​​​​ ID、​​​​​​​quorum vote 的​​​​​​​数​​​​​​​量​​​​​​​和​​​​​​​其​​​​​​​保​​​​​​​护​​​​​​​方​​​​​​​法​​​​​​​。​​​​​​​
  • 保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​ — 显​​​​​​​示​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​。​​​​​​​其​​​​​​​参​​​​​​​数​​​​​​​根​​​​​​​据​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​的​​​​​​​类​​​​​​​型​​​​​​​而​​​​​​​有​​​​​​​所​​​​​​​不​​​​​​​同​​​​​​​。​​​​​​​例​​​​​​​如​​​​​​​,对​​​​​​​于​​​​​​​用​​​​​​​作​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​的​​​​​​​电​​​​​​​源​​​​​​​控​​​​​​​制​​​​​​​器​​​​​​​,群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​定​​​​​​​义​​​​​​​了​​​​​​​电​​​​​​​源​​​​​​​控​​​​​​​制​​​​​​​器​​​​​​​的​​​​​​​名​​​​​​​字​​​​​​​、​​​​​​​IP 地​​​​​​​址​​​​​​​、​​​​​​​登​​​​​​​录​​​​​​​帐​​​​​​​号​​​​​​​和​​​​​​​密​​​​​​​码​​​​​​​。​​​​​​​
  • 被​​​​​​​管​​​​​​​理​​​​​​​的​​​​​​​资​​​​​​​源​​​​​​​ — 显​​​​​​​示​​​​​​​创​​​​​​​建​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​所​​​​​​​必​​​​​​​需​​​​​​​的​​​​​​​资​​​​​​​源​​​​​​​。​​​​​​​被​​​​​​​管​​​​​​​理​​​​​​​的​​​​​​​资​​​​​​​源​​​​​​​包​​​​​​​括​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​的​​​​​​​定​​​​​​​义​​​​​​​、​​​​​​​资​​​​​​​源​​​​​​​(如​​​​​​​ IP 地​​​​​​​址​​​​​​​)和​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​它​​​​​​​也​​​​​​​定​​​​​​​义​​​​​​​了​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​和​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​行​​​​​​​为​​​​​​​。​​​​​​​

1.4. 高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​

高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​提​​​​​​​供​​​​​​​在​​​​​​​ Red Hat 群​​​​​​​集​​​​​​​中​​​​​​​创​​​​​​​建​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​(cluster service)的​​​​​​​能​​​​​​​力​​​​​​​。​​​​​​​Red Hat 群​​​​​​​集​​​​​​​里​​​​​​​的​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​的​​​​​​​关​​​​​​​键​​​​​​​组​​​​​​​件​​​​​​​是​​​​​​​ rgmanager,它​​​​​​​实​​​​​​​现​​​​​​​了​​​​​​​ off-the-shelf 应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​的​​​​​​​冷​​​​​​​失​​​​​​​效​​​​​​​切​​​​​​​换​​​​​​​(cold failover)。​​​​​​​在​​​​​​​ Red Hat 群​​​​​​​集​​​​​​​里​​​​​​​,应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​和​​​​​​​其​​​​​​​他​​​​​​​群​​​​​​​集​​​​​​​资​​​​​​​源​​​​​​​进​​​​​​​行​​​​​​​配​​​​​​​置​​​​​​​来​​​​​​​组​​​​​​​成​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​可​​​​​​​以​​​​​​​从​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​到​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​,而​​​​​​​不​​​​​​​会​​​​​​​对​​​​​​​群​​​​​​​集​​​​​​​客​​​​​​​户​​​​​​​有​​​​​​​明​​​​​​​显​​​​​​​的​​​​​​​影​​​​​​​响​​​​​​​。​​​​​​​当​​​​​​​某​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​发​​​​​​​生​​​​​​​故​​​​​​​障​​​​​​​或​​​​​​​群​​​​​​​集​​​​​​​系​​​​​​​统​​​​​​​管​​​​​​​理​​​​​​​员​​​​​​​将​​​​​​​服​​​​​​​务​​​​​​​迁​​​​​​​移​​​​​​​(例​​​​​​​如​​​​​​​,需​​​​​​​要​​​​​​​对​​​​​​​这​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​进​​​​​​​行​​​​​​​预​​​​​​​定​​​​​​​的​​​​​​​维​​​​​​​护​​​​​​​时​​​​​​​)到​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​时​​​​​​​,群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​都​​​​​​​可​​​​​​​以​​​​​​​发​​​​​​​生​​​​​​​。​​​​​​​
要​​​​​​​创​​​​​​​建​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​,您​​​​​​​必​​​​​​​须​​​​​​​在​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​里​​​​​​​进​​​​​​​行​​​​​​​配​​​​​​​置​​​​​​​。​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​由​​​​​​​群​​​​​​​集​​​​​​​资​​​​​​​源​​​​​​​组​​​​​​​成​​​​​​​。​​​​​​​群​​​​​​​集​​​​​​​资​​​​​​​源​​​​​​​是​​​​​​​您​​​​​​​在​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​里​​​​​​​创​​​​​​​建​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​的​​​​​​​构​​​​​​​建​​​​​​​块​​​​​​​(building block) — 例​​​​​​​如​​​​​​​,IP 地​​​​​​​址​​​​​​​、​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​初​​​​​​​始​​​​​​​化​​​​​​​脚​​​​​​​本​​​​​​​或​​​​​​​者​​​​​​​ Red Hat GFS 共​​​​​​​享​​​​​​​分​​​​​​​区​​​​​​​。​​​​​​​
You can associate a cluster service with a failover domain. A failover domain is a subset of cluster nodes that are eligible to run a particular cluster service (refer to 图 1.9 “故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​”).

注意

故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​不​​​​​​​是​​​​​​​必​​​​​​​需​​​​​​​的​​​​​​​。​​​​​​​
群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​在​​​​​​​某​​​​​​​个​​​​​​​时​​​​​​​间​​​​​​​只​​​​​​​能​​​​​​​在​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​中​​​​​​​运​​​​​​​行​​​​​​​,这​​​​​​​样​​​​​​​可​​​​​​​以​​​​​​​维​​​​​​​护​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​完​​​​​​​整​​​​​​​性​​​​​​​。​​​​​​​您​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​故​​​​​​​障​​​​​​​效​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​里​​​​​​​指​​​​​​​定​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​的​​​​​​​优​​​​​​​先​​​​​​​级​​​​​​​。​​​​​​​指​​​​​​​定​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​的​​​​​​​优​​​​​​​先​​​​​​​级​​​​​​​是​​​​​​​通​​​​​​​过​​​​​​​为​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​里​​​​​​​的​​​​​​​每​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​分​​​​​​​配​​​​​​​优​​​​​​​先​​​​​​​级​​​​​​​别​​​​​​​实​​​​​​​现​​​​​​​的​​​​​​​。​​​​​​​优​​​​​​​先​​​​​​​级​​​​​​​决​​​​​​​定​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​的​​​​​​​顺​​​​​​​序​​​​​​​ — 决​​​​​​​定​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​应​​​​​​​该​​​​​​​切​​​​​​​换​​​​​​​到​​​​​​​哪​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​没​​​​​​​有​​​​​​​指​​​​​​​定​​​​​​​优​​​​​​​先​​​​​​​级​​​​​​​,群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​将​​​​​​​可​​​​​​​能​​​​​​​切​​​​​​​换​​​​​​​至​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​里​​​​​​​的​​​​​​​任​​​​​​​意​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​而​​​​​​​且​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​指​​​​​​​定​​​​​​​是​​​​​​​否​​​​​​​限​​​​​​​制​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​在​​​​​​​其​​​​​​​相​​​​​​​关​​​​​​​的​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​里​​​​​​​的​​​​​​​节​​​​​​​点​​​​​​​中​​​​​​​运​​​​​​​行​​​​​​​。​​​​​​​(如​​​​​​​果​​​​​​​与​​​​​​​非​​​​​​​限​​​​​​​制​​​​​​​性​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​相​​​​​​​关​​​​​​​联​​​​​​​,在​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​里​​​​​​​的​​​​​​​成​​​​​​​员​​​​​​​都​​​​​​​不​​​​​​​可​​​​​​​用​​​​​​​的​​​​​​​时​​​​​​​候​​​​​​​,群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​任​​​​​​​意​​​​​​​节​​​​​​​点​​​​​​​中​​​​​​​启​​​​​​​动​​​​​​​。​​​​​​​)
In 图 1.9 “故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​”, Failover Domain 1 is configured to restrict failover within that domain; therefore, Cluster Service X can only fail over between Node A and Node B. Failover Domain 2 is also configured to restrict failover with its domain; additionally, it is configured for failover priority. Failover Domain 2 priority is configured with Node C as priority 1, Node B as priority 2, and Node D as priority 3. If Node C fails, Cluster Service Y fails over to Node B next. If it cannot fail over to Node B, it tries failing over to Node D. Failover Domain 3 is configured with no priority and no restrictions. If the node that Cluster Service Z is running on fails, Cluster Service Z tries failing over to one of the nodes in Failover Domain 3. However, if none of those nodes is available, Cluster Service Z can fail over to any node in the cluster.
故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​

图 1.9. 故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​

图 1.10 “Web Server Cluster Service Example” shows an example of a high-availability cluster service that is a web server named "content-webserver". It is running in cluster node B and is in a failover domain that consists of nodes A, B, and D. In addition, the failover domain is configured with a failover priority to fail over to node D before node A and to restrict failover to nodes only in that failover domain. The cluster service comprises these cluster resources:
  • IP 地​​​​​​​址​​​​​​​资​​​​​​​源​​​​​​​ — 10.10.10.201。​​​​​​​
  • An application resource named "httpd-content" — a web server application init script /etc/init.d/httpd (specifying httpd).
  • A file system resource — Red Hat GFS named "gfs-content-webserver".
Web Server Cluster Service Example

图 1.10. Web Server Cluster Service Example

客​​​​​​​户​​​​​​​通​​​​​​​过​​​​​​​ IP 地​​​​​​​址​​​​​​​ 10.10.10.201 访​​​​​​​问​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​,和​​​​​​​ web 服​​​​​​​务​​​​​​​器​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​ httpd-content 进​​​​​​​行​​​​​​​交​​​​​​​互​​​​​​​。​​​​​​​httpd-content 使​​​​​​​用​​​​​​​ gfs-content-webserver 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​节​​​​​​​点​​​​​​​ B 发​​​​​​​生​​​​​​​故​​​​​​​障​​​​​​​,content-webserver 群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​将​​​​​​​切​​​​​​​换​​​​​​​到​​​​​​​节​​​​​​​点​​​​​​​ D。​​​​​​​如​​​​​​​果​​​​​​​节​​​​​​​点​​​​​​​ D 不​​​​​​​可​​​​​​​用​​​​​​​或​​​​​​​者​​​​​​​也​​​​​​​发​​​​​​​生​​​​​​​了​​​​​​​故​​​​​​​障​​​​​​​,服​​​​​​​务​​​​​​​将​​​​​​​切​​​​​​​换​​​​​​​至​​​​​​​节​​​​​​​点​​​​​​​ A。​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​将​​​​​​​对​​​​​​​群​​​​​​​集​​​​​​​客​​​​​​​户​​​​​​​不​​​​​​​产​​​​​​​生​​​​​​​明​​​​​​​显​​​​​​​的​​​​​​​影​​​​​​​响​​​​​​​。​​​​​​​和​​​​​​​切​​​​​​​换​​​​​​​前​​​​​​​一​​​​​​​样​​​​​​​,群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​将​​​​​​​可​​​​​​​以​​​​​​​用​​​​​​​相​​​​​​​同​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​从​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​进​​​​​​​行​​​​​​​访​​​​​​​问​​​​​​​。​​​​​​​

1.5. Red Hat GFS

Red Hat GFS 是​​​​​​​允​​​​​​​许​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​同​​​​​​​时​​​​​​​访​​​​​​​问​​​​​​​某​​​​​​​个​​​​​​​共​​​​​​​享​​​​​​​的​​​​​​​块​​​​​​​设​​​​​​​备​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​。​​​​​​​GFS 是​​​​​​​一​​​​​​​个​​​​​​​原​​​​​​​生​​​​​​​的​​​​​​​(native)文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​,它​​​​​​​直​​​​​​​接​​​​​​​和​​​​​​​ Linux 内​​​​​​​核​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​接​​​​​​​口​​​​​​​的​​​​​​​ VFS 层​​​​​​​进​​​​​​​行​​​​​​​通​​​​​​​信​​​​​​​。​​​​​​​GFS 使​​​​​​​用​​​​​​​分​​​​​​​布​​​​​​​式​​​​​​​元​​​​​​​数​​​​​​​据​​​​​​​和​​​​​​​多​​​​​​​个​​​​​​​日​​​​​​​志​​​​​​​来​​​​​​​优​​​​​​​化​​​​​​​群​​​​​​​集​​​​​​​操​​​​​​​作​​​​​​​。​​​​​​​要​​​​​​​维​​​​​​​护​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​完​​​​​​​整​​​​​​​性​​​​​​​,GFS 使​​​​​​​用​​​​​​​一​​​​​​​个​​​​​​​锁​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​来​​​​​​​协​​​​​​​调​​​​​​​ I/O。​​​​​​​当​​​​​​​节​​​​​​​点​​​​​​​修​​​​​​​改​​​​​​​了​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​上​​​​​​​的​​​​​​​数​​​​​​​据​​​​​​​,这​​​​​​​个​​​​​​​改​​​​​​​动​​​​​​​马​​​​​​​上​​​​​​​为​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​其​​​​​​​他​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​可​​​​​​​见​​​​​​​。​​​​​​​
使​​​​​​​用​​​​​​​ Red Hat GFS,您​​​​​​​可​​​​​​​以​​​​​​​从​​​​​​​如​​​​​​​下​​​​​​​几​​​​​​​方​​​​​​​面​​​​​​​受​​​​​​​益​​​​​​​并​​​​​​​实​​​​​​​现​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​的​​​​​​​最​​​​​​​大​​​​​​​正​​​​​​​常​​​​​​​运​​​​​​​行​​​​​​​时​​​​​​​间​​​​​​​:
  • 简​​​​​​​化​​​​​​​数​​​​​​​据​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​
    • 整​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​只​​​​​​​需​​​​​​​安​​​​​​​装​​​​​​​一​​​​​​​次​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​或​​​​​​​补​​​​​​​丁​​​​​​​
    • 消​​​​​​​除​​​​​​​对​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​冗​​​​​​​余​​​​​​​拷​​​​​​​贝​​​​​​​(复​​​​​​​制​​​​​​​)的​​​​​​​需​​​​​​​要​​​​​​​。​​​​​​​
    • 启​​​​​​​用​​​​​​​多​​​​​​​个​​​​​​​客​​​​​​​户​​​​​​​对​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​并​​​​​​​行​​​​​​​读​​​​​​​/写​​​​​​​操​​​​​​​作​​​​​​​。​​​​​​​
    • 简​​​​​​​化​​​​​​​备​​​​​​​份​​​​​​​和​​​​​​​恢​​​​​​​复​​​​​​​(只​​​​​​​需​​​​​​​要​​​​​​​备​​​​​​​份​​​​​​​或​​​​​​​恢​​​​​​​复​​​​​​​一​​​​​​​个​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​)。​​​​​​​
  • 最​​​​​​​大​​​​​​​化​​​​​​​对​​​​​​​存​​​​​​​储​​​​​​​资​​​​​​​源​​​​​​​的​​​​​​​使​​​​​​​用​​​​​​​;最​​​​​​​小​​​​​​​化​​​​​​​存​​​​​​​储​​​​​​​的​​​​​​​管​​​​​​​理​​​​​​​开​​​​​​​销​​​​​​​。​​​​​​​
    • 将​​​​​​​存​​​​​​​储​​​​​​​作​​​​​​​为​​​​​​​一​​​​​​​个​​​​​​​整​​​​​​​体​​​​​​​而​​​​​​​不​​​​​​​是​​​​​​​根​​​​​​​据​​​​​​​分​​​​​​​区​​​​​​​管​​​​​​​理​​​​​​​。​​​​​​​
    • 通​​​​​​​过​​​​​​​消​​​​​​​除​​​​​​​数​​​​​​​据​​​​​​​复​​​​​​​制​​​​​​​的​​​​​​​需​​​​​​​要​​​​​​​从​​​​​​​而​​​​​​​减​​​​​​​少​​​​​​​总​​​​​​​体​​​​​​​的​​​​​​​存​​​​​​​储​​​​​​​需​​​​​​​求​​​​​​​。​​​​​​​
  • 通​​​​​​​过​​​​​​​即​​​​​​​时​​​​​​​添​​​​​​​加​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​或​​​​​​​存​​​​​​​储​​​​​​​空​​​​​​​间​​​​​​​来​​​​​​​无​​​​​​​缝​​​​​​​地​​​​​​​扩​​​​​​​展​​​​​​​群​​​​​​​集​​​​​​​。​​​​​​​
    • 不​​​​​​​需​​​​​​​要​​​​​​​复​​​​​​​杂​​​​​​​的​​​​​​​技​​​​​​​术​​​​​​​来​​​​​​​对​​​​​​​存​​​​​​​储​​​​​​​空​​​​​​​间​​​​​​​分​​​​​​​区​​​​​​​。​​​​​​​
    • 通​​​​​​​过​​​​​​​挂​​​​​​​载​​​​​​​至​​​​​​​普​​​​​​​通​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​来​​​​​​​实​​​​​​​时​​​​​​​在​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​添​​​​​​​加​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​
运​​​​​​​行​​​​​​​ Red Hat GFS 的​​​​​​​节​​​​​​​点​​​​​​​都​​​​​​​用​​​​​​​ Red Hat Cluster Suite 配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​进​​​​​​​行​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​。​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​通​​​​​​​过​​​​​​​群​​​​​​​集​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​(Cluster Logical Volume Manager,CLVM)进​​​​​​​行​​​​​​​管​​​​​​​理​​​​​​​。​​​​​​​Red Hat GFS 在​​​​​​​ Red Hat 群​​​​​​​集​​​​​​​的​​​​​​​ GFS 节​​​​​​​点​​​​​​​间​​​​​​​提​​​​​​​供​​​​​​​数​​​​​​​据​​​​​​​共​​​​​​​享​​​​​​​。​​​​​​​GFS 提​​​​​​​供​​​​​​​ Red Hat 群​​​​​​​集​​​​​​​的​​​​​​​ GFS 节​​​​​​​点​​​​​​​的​​​​​​​单​​​​​​​一​​​​​​​的​​​​​​​、​​​​​​​一​​​​​​​致​​​​​​​的​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​命​​​​​​​名​​​​​​​空​​​​​​​间​​​​​​​的​​​​​​​视​​​​​​​图​​​​​​​。​​​​​​​GFS 允​​​​​​​许​​​​​​​安​​​​​​​装​​​​​​​和​​​​​​​运​​​​​​​行​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​而​​​​​​​无​​​​​​​需​​​​​​​了​​​​​​​解​​​​​​​底​​​​​​​层​​​​​​​的​​​​​​​存​​​​​​​储​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​。​​​​​​​而​​​​​​​且​​​​​​​,GFS 提​​​​​​​供​​​​​​​通​​​​​​​常​​​​​​​在​​​​​​​企​​​​​​​业​​​​​​​环​​​​​​​境​​​​​​​才​​​​​​​要​​​​​​​求​​​​​​​的​​​​​​​功​​​​​​​能​​​​​​​,如​​​​​​​配​​​​​​​额​​​​​​​(quota)、​​​​​​​多​​​​​​​重​​​​​​​日​​​​​​​志​​​​​​​和​​​​​​​多​​​​​​​路​​​​​​​径​​​​​​​(multipath)支​​​​​​​持​​​​​​​。​​​​​​​
GFS 根​​​​​​​据​​​​​​​存​​​​​​​储​​​​​​​环​​​​​​​境​​​​​​​的​​​​​​​性​​​​​​​能​​​​​​​、​​​​​​​可​​​​​​​伸​​​​​​​缩​​​​​​​性​​​​​​​和​​​​​​​经​​​​​​​济​​​​​​​性​​​​​​​需​​​​​​​要​​​​​​​来​​​​​​​提​​​​​​​供​​​​​​​网​​​​​​​络​​​​​​​存​​​​​​​储​​​​​​​的​​​​​​​通​​​​​​​用​​​​​​​方​​​​​​​法​​​​​​​。​​​​​​​本​​​​​​​章​​​​​​​提​​​​​​​供​​​​​​​一​​​​​​​些​​​​​​​非​​​​​​​常​​​​​​​基​​​​​​​础​​​​​​​的​​​​​​​、​​​​​​​简​​​​​​​略​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​来​​​​​​​帮​​​​​​​助​​​​​​​您​​​​​​​理​​​​​​​解​​​​​​​ GFS。​​​​​​​
You can deploy GFS in a variety of configurations to suit your needs for performance, scalability, and economy. For superior performance and scalability, you can deploy GFS in a cluster that is connected directly to a SAN. For more economical needs, you can deploy GFS in a cluster that is connected to a LAN with servers that use GNBD (Global Network Block Device) or to iSCSI (Internet Small Computer System Interface) devices. (For more information about GNBD, refer to 第 1.7 节 “全​​​​​​​球​​​​​​​网​​​​​​​络​​​​​​​块​​​​​​​设​​​​​​​备​​​​​​​”.)
下​​​​​​​面​​​​​​​的​​​​​​​章​​​​​​​节​​​​​​​提​​​​​​​供​​​​​​​的​​​​​​​例​​​​​​​子​​​​​​​,展​​​​​​​示​​​​​​​了​​​​​​​怎​​​​​​​样​​​​​​​部​​​​​​​署​​​​​​​ GFS 来​​​​​​​满​​​​​​​足​​​​​​​您​​​​​​​对​​​​​​​性​​​​​​​能​​​​​​​、​​​​​​​灵​​​​​​​活​​​​​​​性​​​​​​​和​​​​​​​经​​​​​​​济​​​​​​​性​​​​​​​的​​​​​​​需​​​​​​​要​​​​​​​:

注意

GFS 部​​​​​​​署​​​​​​​示​​​​​​​例​​​​​​​只​​​​​​​是​​​​​​​演​​​​​​​示​​​​​​​了​​​​​​​基​​​​​​​本​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​;在​​​​​​​实​​​​​​​践​​​​​​​中​​​​​​​,您​​​​​​​可​​​​​​​能​​​​​​​需​​​​​​​要​​​​​​​将​​​​​​​这​​​​​​​些​​​​​​​配​​​​​​​置​​​​​​​进​​​​​​​行​​​​​​​组​​​​​​​合​​​​​​​。​​​​​​​

1.5.1. 出​​​​​​​众​​​​​​​的​​​​​​​性​​​​​​​能​​​​​​​和​​​​​​​灵​​​​​​​活​​​​​​​性​​​​​​​

You can obtain the highest shared-file performance when applications access storage directly. The GFS SAN configuration in 图 1.11 “GFS with a SAN” provides superior file performance for shared files and file systems. Linux applications run directly on cluster nodes using GFS. Without file protocols or storage servers to slow data access, performance is similar to individual Linux servers with directly connected storage; yet, each GFS application node has equal access to all data files. GFS supports over 300 GFS nodes.
GFS with a SAN

图 1.11. GFS with a SAN

1.5.2. 性​​​​​​​能​​​​​​​、​​​​​​​灵​​​​​​​活​​​​​​​性​​​​​​​、​​​​​​​合​​​​​​​适​​​​​​​的​​​​​​​价​​​​​​​格​​​​​​​

Multiple Linux client applications on a LAN can share the same SAN-based data as shown in 图 1.12 “GFS and GNBD with a SAN”. SAN block storage is presented to network clients as block storage devices by GNBD servers. From the perspective of a client application, storage is accessed as if it were directly attached to the server in which the application is running. Stored data is actually on the SAN. Storage devices and data can be equally shared by network client applications. File locking and sharing functions are handled by GFS for each network client.
GFS and GNBD with a SAN

图 1.12. GFS and GNBD with a SAN

1.5.3. 经​​​​​​​济​​​​​​​性​​​​​​​和​​​​​​​性​​​​​​​能​​​​​​​

图 1.13 “使​​​​​​​用​​​​​​​直​​​​​​​连​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​的​​​​​​​ GFS 和​​​​​​​ GNBD ” shows how Linux client applications can take advantage of an existing Ethernet topology to gain shared access to all block storage devices. Client data files and file systems can be shared with GFS on each client. Application failover can be fully automated with Red Hat Cluster Suite.
使​​​​​​​用​​​​​​​直​​​​​​​连​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​的​​​​​​​ GFS 和​​​​​​​ GNBD

图 1.13. 使​​​​​​​用​​​​​​​直​​​​​​​连​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​的​​​​​​​ GFS 和​​​​​​​ GNBD

1.6. 群​​​​​​​集​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​

群​​​​​​​集​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​(Cluster Logical Volume Manager,CLVM)提​​​​​​​供​​​​​​​了​​​​​​​ LVM2 的​​​​​​​群​​​​​​​集​​​​​​​版​​​​​​​本​​​​​​​。​​​​​​​CLVM 具​​​​​​​备​​​​​​​和​​​​​​​单​​​​​​​节​​​​​​​点​​​​​​​上​​​​​​​的​​​​​​​ LVM2 相​​​​​​​同​​​​​​​的​​​​​​​能​​​​​​​力​​​​​​​,但​​​​​​​它​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​为​​​​​​​ Red Hat 群​​​​​​​集​​​​​​​里​​​​​​​的​​​​​​​所​​​​​​​有​​​​​​​节​​​​​​​点​​​​​​​所​​​​​​​用​​​​​​​。​​​​​​​用​​​​​​​ CLVM 创​​​​​​​建​​​​​​​的​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​可​​​​​​​为​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​的​​​​​​​所​​​​​​​有​​​​​​​节​​​​​​​点​​​​​​​所​​​​​​​用​​​​​​​。​​​​​​​
The key component in CLVM is clvmd. clvmd is a daemon that provides clustering extensions to the standard LVM2 tool set and allows LVM2 commands to manage shared storage. clvmd runs in each cluster node and distributes LVM metadata updates in a cluster, thereby presenting each cluster node with the same view of the logical volumes (refer to 图 1.14 “CLVM Overview”). Logical volumes created with CLVM on shared storage are visible to all nodes that have access to the shared storage. CLVM allows a user to configure logical volumes on shared storage by locking access to physical storage while a logical volume is being configured. CLVM uses the lock-management service provided by the cluster infrastructure (refer to 第 1.3 节 “Cluster Infrastructure”).

注意

在​​​​​​​红​​​​​​​帽​​​​​​​群​​​​​​​集​​​​​​​套​​​​​​​件​​​​​​​中​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​共​​​​​​​享​​​​​​​存​​​​​​​储​​​​​​​要​​​​​​​求​​​​​​​您​​​​​​​运​​​​​​​行​​​​​​​群​​​​​​​集​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​(clvmd)或​​​​​​​者​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​代​​​​​​​理​​​​​​​(HA-LVM)。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​无​​​​​​​法​​​​​​​使​​​​​​​用​​​​​​​ clvmd 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​或​​​​​​​者​​​​​​​ HA-LVM,无​​​​​​​论​​​​​​​是​​​​​​​因​​​​​​​为​​​​​​​操​​​​​​​作​​​​​​​原​​​​​​​因​​​​​​​还​​​​​​​是​​​​​​​您​​​​​​​没​​​​​​​有​​​​​​​正​​​​​​​确​​​​​​​的​​​​​​​权​​​​​​​利​​​​​​​,您​​​​​​​就​​​​​​​一​​​​​​​定​​​​​​​不​​​​​​​能​​​​​​​在​​​​​​​共​​​​​​​享​​​​​​​磁​​​​​​​盘​​​​​​​中​​​​​​​使​​​​​​​用​​​​​​​单​​​​​​​一​​​​​​​事​​​​​​​件​​​​​​​ LVM,因​​​​​​​为​​​​​​​这​​​​​​​将​​​​​​​导​​​​​​​致​​​​​​​数​​​​​​​据​​​​​​​崩​​​​​​​溃​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​有​​​​​​​任​​​​​​​何​​​​​​​疑​​​​​​​问​​​​​​​,请​​​​​​​联​​​​​​​络​​​​​​​您​​​​​​​的​​​​​​​红​​​​​​​帽​​​​​​​服​​​​​​​务​​​​​​​代​​​​​​​表​​​​​​​。​​​​​​​

注意

使​​​​​​​用​​​​​​​ CLVM 需​​​​​​​要​​​​​​​对​​​​​​​ /etc/lvm/lvm.conf 进​​​​​​​行​​​​​​​群​​​​​​​集​​​​​​​范​​​​​​​围​​​​​​​的​​​​​​​锁​​​​​​​定​​​​​​​方​​​​​​​面​​​​​​​的​​​​​​​少​​​​​​​许​​​​​​​修​​​​​​​改​​​​​​​。​​​​​​​
CLVM Overview

图 1.14. CLVM Overview

You can configure CLVM using the same commands as LVM2, using the LVM graphical user interface (refer to 图 1.15 “LVM Graphical User Interface”), or using the storage configuration function of the Conga cluster configuration graphical user interface (refer to 图 1.16 “Conga LVM Graphical User Interface”) . 图 1.17 “Creating Logical Volumes” shows the basic concept of creating logical volumes from Linux partitions and shows the commands used to create logical volumes.
LVM Graphical User Interface

图 1.15. LVM Graphical User Interface

Conga LVM Graphical User Interface

图 1.16. Conga LVM Graphical User Interface

Creating Logical Volumes

图 1.17. Creating Logical Volumes

1.7. 全​​​​​​​球​​​​​​​网​​​​​​​络​​​​​​​块​​​​​​​设​​​​​​​备​​​​​​​

全​​​​​​​球​​​​​​​网​​​​​​​络​​​​​​​块​​​​​​​设​​​​​​​备​​​​​​​(Global Network Block Device,GNBD)提​​​​​​​供​​​​​​​通​​​​​​​过​​​​​​​ TCP/IP 访​​​​​​​问​​​​​​​ Red Hat GFS 块​​​​​​​设​​​​​​​备​​​​​​​的​​​​​​​功​​​​​​​能​​​​​​​。​​​​​​​GNBD 从​​​​​​​概​​​​​​​念​​​​​​​上​​​​​​​和​​​​​​​ NBD 相​​​​​​​似​​​​​​​;然​​​​​​​而​​​​​​​,GNBD 是​​​​​​​和​​​​​​​ GFS 相​​​​​​​关​​​​​​​的​​​​​​​且​​​​​​​为​​​​​​​ GFS 单​​​​​​​独​​​​​​​进​​​​​​​行​​​​​​​了​​​​​​​调​​​​​​​整​​​​​​​。​​​​​​​当​​​​​​​需​​​​​​​要​​​​​​​更​​​​​​​健​​​​​​​壮​​​​​​​的​​​​​​​技​​​​​​​术​​​​​​​时​​​​​​​,GNBD 是​​​​​​​有​​​​​​​用​​​​​​​处​​​​​​​的​​​​​​​ — 光​​​​​​​纤​​​​​​​通​​​​​​​道​​​​​​​或​​​​​​​ single-initiator SCSI — 不​​​​​​​是​​​​​​​必​​​​​​​需​​​​​​​的​​​​​​​或​​​​​​​成​​​​​​​本​​​​​​​过​​​​​​​高​​​​​​​。​​​​​​​
GNBD consists of two major components: a GNBD client and a GNBD server. A GNBD client runs in a node with GFS and imports a block device exported by a GNBD server. A GNBD server runs in another node and exports block-level storage from its local storage (either directly attached storage or SAN storage). Refer to 图 1.18 “GNBD 概​​​​​​​述​​​​​​​”. Multiple GNBD clients can access a device exported by a GNBD server, thus making a GNBD suitable for use by a group of nodes running GFS.
GNBD 概​​​​​​​述​​​​​​​

图 1.18. GNBD 概​​​​​​​述​​​​​​​

1.8. Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​

Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​(LVS)是​​​​​​​一​​​​​​​组​​​​​​​集​​​​​​​成​​​​​​​的​​​​​​​软​​​​​​​件​​​​​​​组​​​​​​​件​​​​​​​,它​​​​​​​用​​​​​​​于​​​​​​​在​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​之​​​​​​​间​​​​​​​平​​​​​​​衡​​​​​​​ IP 负​​​​​​​载​​​​​​​。​​​​​​​LVS 运​​​​​​​行​​​​​​​在​​​​​​​一​​​​​​​对​​​​​​​配​​​​​​​置​​​​​​​相​​​​​​​同​​​​​​​的​​​​​​​机​​​​​​​器​​​​​​​上​​​​​​​:一​​​​​​​个​​​​​​​作​​​​​​​为​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​而​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​作​​​​​​​为​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​。​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​扮​​​​​​​演​​​​​​​两​​​​​​​个​​​​​​​角​​​​​​​色​​​​​​​:
  • 在​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​之​​​​​​​间​​​​​​​平​​​​​​​衡​​​​​​​负​​​​​​​载​​​​​​​。​​​​​​​
  • 检​​​​​​​查​​​​​​​每​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​完​​​​​​​整​​​​​​​性​​​​​​​。​​​​​​​
备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​监​​​​​​​控​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​并​​​​​​​在​​​​​​​它​​​​​​​发​​​​​​​生​​​​​​​故​​​​​​​障​​​​​​​时​​​​​​​进​​​​​​​行​​​​​​​接​​​​​​​管​​​​​​​。​​​​​​​
图 1.19 “Components of a Running LVS Cluster” provides an overview of the LVS components and their interrelationship.
Components of a Running LVS Cluster

图 1.19. Components of a Running LVS Cluster

pulse 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​运​​​​​​​行​​​​​​​在​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​以​​​​​​​及​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​上​​​​​​​。​​​​​​​在​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​上​​​​​​​,pulse 发​​​​​​​送​​​​​​​ heartbeat 到​​​​​​​活​​​​​​​动​​​​​​​路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​公​​​​​​​用​​​​​​​接​​​​​​​口​​​​​​​,确​​​​​​​保​​​​​​​活​​​​​​​动​​​​​​​路​​​​​​​由​​​​​​​器​​​​​​​正​​​​​​​常​​​​​​​工​​​​​​​作​​​​​​​。​​​​​​​在​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​上​​​​​​​,pulse 启​​​​​​​动​​​​​​​ lvs 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​并​​​​​​​响​​​​​​​应​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​ heartbeat 查​​​​​​​询​​​​​​​。​​​​​​​
一​​​​​​​旦​​​​​​​启​​​​​​​动​​​​​​​,lvs 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​将​​​​​​​调​​​​​​​用​​​​​​​ ipvsadm 工​​​​​​​具​​​​​​​来​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​维​​​​​​​护​​​​​​​内​​​​​​​核​​​​​​​里​​​​​​​的​​​​​​​ IPVS(IP 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​)路​​​​​​​由​​​​​​​表​​​​​​​,并​​​​​​​为​​​​​​​每​​​​​​​台​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​配​​​​​​​置​​​​​​​的​​​​​​​的​​​​​​​每​​​​​​​个​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​启​​​​​​​动​​​​​​​一​​​​​​​个​​​​​​​ nanny 进​​​​​​​程​​​​​​​。​​​​​​​每​​​​​​​个​​​​​​​ nanny 进​​​​​​​程​​​​​​​检​​​​​​​查​​​​​​​一​​​​​​​台​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​的​​​​​​​一​​​​​​​个​​​​​​​配​​​​​​​置​​​​​​​好​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​状​​​​​​​态​​​​​​​,并​​​​​​​告​​​​​​​诉​​​​​​​ lvs 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​这​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​是​​​​​​​否​​​​​​​发​​​​​​​生​​​​​​​了​​​​​​​故​​​​​​​障​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​检​​​​​​​测​​​​​​​到​​​​​​​了​​​​​​​故​​​​​​​障​​​​​​​,lvs 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​将​​​​​​​指​​​​​​​引​​​​​​​ ipvsadm 从​​​​​​​ IPVS 路​​​​​​​由​​​​​​​表​​​​​​​中​​​​​​​删​​​​​​​除​​​​​​​那​​​​​​​台​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​条​​​​​​​目​​​​​​​。​​​​​​​
如​​​​​​​果​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​没​​​​​​​有​​​​​​​从​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​收​​​​​​​到​​​​​​​响​​​​​​​应​​​​​​​,它​​​​​​​将​​​​​​​初​​​​​​​始​​​​​​​化​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​,调​​​​​​​用​​​​​​​ send_arp 将​​​​​​​所​​​​​​​有​​​​​​​虚​​​​​​​拟​​​​​​​ IP 地​​​​​​​址​​​​​​​重​​​​​​​新​​​​​​​分​​​​​​​配​​​​​​​给​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​ NIC 硬​​​​​​​件​​​​​​​地​​​​​​​址​​​​​​​(MAC 地​​​​​​​址​​​​​​​),并​​​​​​​通​​​​​​​过​​​​​​​公​​​​​​​用​​​​​​​以​​​​​​​及​​​​​​​专​​​​​​​用​​​​​​​网​​​​​​​络​​​​​​​接​​​​​​​口​​​​​​​向​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​发​​​​​​​送​​​​​​​命​​​​​​​令​​​​​​​来​​​​​​​关​​​​​​​闭​​​​​​​其​​​​​​​ lvs 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​,然​​​​​​​后​​​​​​​在​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​上​​​​​​​启​​​​​​​动​​​​​​​ lvs 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​来​​​​​​​接​​​​​​​受​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​。​​​​​​​
对​​​​​​​于​​​​​​​访​​​​​​​问​​​​​​​宿​​​​​​​主​​​​​​​服​​​​​​​务​​​​​​​(如​​​​​​​网​​​​​​​站​​​​​​​或​​​​​​​数​​​​​​​据​​​​​​​库​​​​​​​应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​)的​​​​​​​外​​​​​​​部​​​​​​​用​​​​​​​户​​​​​​​,LVS 就​​​​​​​象​​​​​​​是​​​​​​​一​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​然​​​​​​​而​​​​​​​,用​​​​​​​户​​​​​​​实​​​​​​​际​​​​​​​上​​​​​​​访​​​​​​​问​​​​​​​了​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​后​​​​​​​面​​​​​​​的​​​​​​​多​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​
因​​​​​​​为​​​​​​​在​​​​​​​ LVS 里​​​​​​​没​​​​​​​有​​​​​​​内​​​​​​​置​​​​​​​组​​​​​​​件​​​​​​​在​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​间​​​​​​​共​​​​​​​享​​​​​​​数​​​​​​​据​​​​​​​,您​​​​​​​有​​​​​​​两​​​​​​​个​​​​​​​基​​​​​​​本​​​​​​​选​​​​​​​择​​​​​​​:
  • 在​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​间​​​​​​​对​​​​​​​数​​​​​​​据​​​​​​​进​​​​​​​行​​​​​​​同​​​​​​​步​​​​​​​。​​​​​​​
  • 在​​​​​​​拓​​​​​​​扑​​​​​​​结​​​​​​​构​​​​​​​中​​​​​​​添​​​​​​​加​​​​​​​用​​​​​​​于​​​​​​​共​​​​​​​享​​​​​​​数​​​​​​​据​​​​​​​访​​​​​​​问​​​​​​​的​​​​​​​第​​​​​​​三​​​​​​​层​​​​​​​。​​​​​​​
对​​​​​​​于​​​​​​​不​​​​​​​允​​​​​​​许​​​​​​​大​​​​​​​量​​​​​​​用​​​​​​​户​​​​​​​在​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​上​​​​​​​载​​​​​​​或​​​​​​​修​​​​​​​改​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​来​​​​​​​说​​​​​​​,第​​​​​​​一​​​​​​​个​​​​​​​选​​​​​​​择​​​​​​​是​​​​​​​首​​​​​​​选​​​​​​​项​​​​​​​。​​​​​​​而​​​​​​​如​​​​​​​果​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​允​​​​​​​许​​​​​​​大​​​​​​​量​​​​​​​用​​​​​​​户​​​​​​​修​​​​​​​改​​​​​​​数​​​​​​​据​​​​​​​,如​​​​​​​电​​​​​​​子​​​​​​​商​​​​​​​务​​​​​​​网​​​​​​​站​​​​​​​,添​​​​​​​加​​​​​​​第​​​​​​​三​​​​​​​层​​​​​​​则​​​​​​​更​​​​​​​为​​​​​​​可​​​​​​​取​​​​​​​。​​​​​​​
有​​​​​​​许​​​​​​​多​​​​​​​方​​​​​​​法​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​间​​​​​​​进​​​​​​​行​​​​​​​数​​​​​​​据​​​​​​​同​​​​​​​步​​​​​​​。​​​​​​​例​​​​​​​如​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​ shell 脚​​​​​​​本​​​​​​​同​​​​​​​时​​​​​​​将​​​​​​​更​​​​​​​新​​​​​​​的​​​​​​​ web 页​​​​​​​面​​​​​​​登​​​​​​​记​​​​​​​到​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​或​​​​​​​者​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​ rsync 等​​​​​​​程​​​​​​​序​​​​​​​以​​​​​​​设​​​​​​​定​​​​​​​的​​​​​​​时​​​​​​​间​​​​​​​间​​​​​​​隔​​​​​​​在​​​​​​​所​​​​​​​有​​​​​​​节​​​​​​​点​​​​​​​间​​​​​​​复​​​​​​​制​​​​​​​改​​​​​​​动​​​​​​​的​​​​​​​数​​​​​​​据​​​​​​​。​​​​​​​然​​​​​​​而​​​​​​​,在​​​​​​​用​​​​​​​户​​​​​​​频​​​​​​​繁​​​​​​​地​​​​​​​上​​​​​​​载​​​​​​​文​​​​​​​件​​​​​​​或​​​​​​​执​​​​​​​行​​​​​​​数​​​​​​​据​​​​​​​库​​​​​​​事​​​​​​​务​​​​​​​的​​​​​​​环​​​​​​​境​​​​​​​里​​​​​​​,使​​​​​​​用​​​​​​​脚​​​​​​​本​​​​​​​或​​​​​​​ rsync 命​​​​​​​令​​​​​​​来​​​​​​​进​​​​​​​行​​​​​​​同​​​​​​​步​​​​​​​并​​​​​​​不​​​​​​​是​​​​​​​最​​​​​​​理​​​​​​​想​​​​​​​的​​​​​​​方​​​​​​​法​​​​​​​。​​​​​​​因​​​​​​​此​​​​​​​,对​​​​​​​于​​​​​​​执​​​​​​​行​​​​​​​大​​​​​​​量​​​​​​​上​​​​​​​载​​​​​​​、​​​​​​​数​​​​​​​据​​​​​​​库​​​​​​​事​​​​​​​务​​​​​​​或​​​​​​​具​​​​​​​有​​​​​​​类​​​​​​​似​​​​​​​流​​​​​​​量​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​,三​​​​​​​层​​​​​​​的​​​​​​​拓​​​​​​​扑​​​​​​​结​​​​​​​构​​​​​​​(three-tiered topology) 更​​​​​​​适​​​​​​​合​​​​​​​于​​​​​​​数​​​​​​​据​​​​​​​同​​​​​​​步​​​​​​​。​​​​​​​

1.8.1. Two-Tier LVS Topology

图 1.20 “Two-Tier LVS Topology” shows a simple LVS configuration consisting of two tiers: LVS routers and real servers. The LVS-router tier consists of one active LVS router and one backup LVS router. The real-server tier consists of real servers connected to the private network. Each LVS router has two network interfaces: one connected to a public network (Internet) and one connected to a private network. A network interface connected to each network allows the LVS routers to regulate traffic between clients on the public network and the real servers on the private network. In 图 1.20 “Two-Tier LVS Topology”, the active LVS router uses Network Address Translation (NAT) to direct traffic from the public network to real servers on the private network, which in turn provide services as requested. The real servers pass all public traffic through the active LVS router. From the perspective of clients on the public network, the LVS router appears as one entity.
Two-Tier LVS Topology

图 1.20. Two-Tier LVS Topology

到​​​​​​​达​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​请​​​​​​​求​​​​​​​被​​​​​​​指​​​​​​​引​​​​​​​到​​​​​​​虚​​​​​​​拟​​​​​​​ IP 地​​​​​​​址​​​​​​​或​​​​​​​ VIP。​​​​​​​这​​​​​​​是​​​​​​​一​​​​​​​个​​​​​​​公​​​​​​​用​​​​​​​的​​​​​​​地​​​​​​​址​​​​​​​,站​​​​​​​点​​​​​​​管​​​​​​​理​​​​​​​员​​​​​​​将​​​​​​​它​​​​​​​和​​​​​​​全​​​​​​​限​​​​​​​定​​​​​​​域​​​​​​​名​​​​​​​相​​​​​​​关​​​​​​​联​​​​​​​,如​​​​​​​ www.example.com,并​​​​​​​分​​​​​​​配​​​​​​​给​​​​​​​一​​​​​​​个​​​​​​​或​​​​​​​多​​​​​​​个​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​[1]。​​​​​​​注​​​​​​​意​​​​​​​,进​​​​​​​行​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​时​​​​​​​,VIP 地​​​​​​​址​​​​​​​将​​​​​​​从​​​​​​​一​​​​​​​个​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​移​​​​​​​植​​​​​​​到​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​,并​​​​​​​保​​​​​​​持​​​​​​​这​​​​​​​个​​​​​​​ IP 地​​​​​​​址​​​​​​​,这​​​​​​​被​​​​​​​称​​​​​​​为​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​(floating IP address)。​​​​​​​
VIP 地​​​​​​​址​​​​​​​可​​​​​​​以​​​​​​​以​​​​​​​连​​​​​​​接​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​到​​​​​​​公​​​​​​​用​​​​​​​网​​​​​​​络​​​​​​​的​​​​​​​接​​​​​​​口​​​​​​​为​​​​​​​别​​​​​​​名​​​​​​​。​​​​​​​例​​​​​​​如​​​​​​​,如​​​​​​​果​​​​​​​ eth0 连​​​​​​​接​​​​​​​至​​​​​​​ Internet,那​​​​​​​么​​​​​​​多​​​​​​​个​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​别​​​​​​​名​​​​​​​ eth0:1。​​​​​​​或​​​​​​​者​​​​​​​,每​​​​​​​个​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​可​​​​​​​以​​​​​​​按​​​​​​​照​​​​​​​服​​​​​​​务​​​​​​​和​​​​​​​单​​​​​​​独​​​​​​​的​​​​​​​设​​​​​​​备​​​​​​​相​​​​​​​关​​​​​​​联​​​​​​​。​​​​​​​例​​​​​​​如​​​​​​​,HTTP 通​​​​​​​信​​​​​​​在​​​​​​​ eth0:1 上​​​​​​​处​​​​​​​理​​​​​​​,而​​​​​​​ FTP 通​​​​​​​信​​​​​​​在​​​​​​​ eth0:2 上​​​​​​​进​​​​​​​行​​​​​​​处​​​​​​​理​​​​​​​。​​​​​​​
在​​​​​​​某​​​​​​​一​​​​​​​时​​​​​​​刻​​​​​​​只​​​​​​​能​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​处​​​​​​​于​​​​​​​活​​​​​​​动​​​​​​​状​​​​​​​态​​​​​​​。​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​角​​​​​​​色​​​​​​​是​​​​​​​将​​​​​​​服​​​​​​​务​​​​​​​请​​​​​​​求​​​​​​​从​​​​​​​虚​​​​​​​拟​​​​​​​ IP 地​​​​​​​址​​​​​​​重​​​​​​​定​​​​​​​向​​​​​​​到​​​​​​​实​​​​​​​际​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​这​​​​​​​种​​​​​​​重​​​​​​​定​​​​​​​向​​​​​​​是​​​​​​​基​​​​​​​于​​​​​​​下​​​​​​​面​​​​​​​ 8 个​​​​​​​负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​算​​​​​​​法​​​​​​​中​​​​​​​的​​​​​​​一​​​​​​​个​​​​​​​:
  • 轮​​​​​​​询​​​​​​​调​​​​​​​度​​​​​​​(Round-Robin Scheduling)— 将​​​​​​​每​​​​​​​个​​​​​​​请​​​​​​​求​​​​​​​按​​​​​​​顺​​​​​​​序​​​​​​​地​​​​​​​在​​​​​​​一​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​池​​​​​​​里​​​​​​​进​​​​​​​行​​​​​​​分​​​​​​​发​​​​​​​。​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​算​​​​​​​法​​​​​​​,所​​​​​​​有​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​都​​​​​​​被​​​​​​​一​​​​​​​视​​​​​​​同​​​​​​​仁​​​​​​​,而​​​​​​​不​​​​​​​考​​​​​​​虑​​​​​​​其​​​​​​​能​​​​​​​力​​​​​​​或​​​​​​​负​​​​​​​载​​​​​​​。​​​​​​​
  • 加​​​​​​​权​​​​​​​轮​​​​​​​询​​​​​​​调​​​​​​​度​​​​​​​(Weighted Round-Robin Scheduling)— 将​​​​​​​每​​​​​​​个​​​​​​​请​​​​​​​求​​​​​​​按​​​​​​​顺​​​​​​​序​​​​​​​地​​​​​​​在​​​​​​​一​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​池​​​​​​​里​​​​​​​进​​​​​​​行​​​​​​​分​​​​​​​发​​​​​​​,但​​​​​​​赋​​​​​​​予​​​​​​​能​​​​​​​力​​​​​​​更​​​​​​​强​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​更​​​​​​​多​​​​​​​的​​​​​​​负​​​​​​​载​​​​​​​。​​​​​​​能​​​​​​​力​​​​​​​用​​​​​​​用​​​​​​​户​​​​​​​分​​​​​​​配​​​​​​​的​​​​​​​权​​​​​​​重​​​​​​​来​​​​​​​表​​​​​​​示​​​​​​​,然​​​​​​​后​​​​​​​根​​​​​​​据​​​​​​​负​​​​​​​载​​​​​​​信​​​​​​​息​​​​​​​动​​​​​​​态​​​​​​​地​​​​​​​进​​​​​​​行​​​​​​​上​​​​​​​下​​​​​​​调​​​​​​​整​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​池​​​​​​​里​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​能​​​​​​​力​​​​​​​区​​​​​​​别​​​​​​​很​​​​​​​大​​​​​​​,这​​​​​​​是​​​​​​​首​​​​​​​选​​​​​​​的​​​​​​​办​​​​​​​法​​​​​​​。​​​​​​​然​​​​​​​而​​​​​​​,如​​​​​​​果​​​​​​​请​​​​​​​求​​​​​​​负​​​​​​​载​​​​​​​突​​​​​​​然​​​​​​​发​​​​​​​生​​​​​​​变​​​​​​​化​​​​​​​,权​​​​​​​重​​​​​​​更​​​​​​​大​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​可​​​​​​​能​​​​​​​会​​​​​​​处​​​​​​​理​​​​​​​超​​​​​​​过​​​​​​​其​​​​​​​份​​​​​​​额​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​。​​​​​​​
  • 最​​​​​​​小​​​​​​​连​​​​​​​接​​​​​​​调​​​​​​​度​​​​​​​(Least-Connection Scheduling)— 给​​​​​​​活​​​​​​​动​​​​​​​连​​​​​​​接​​​​​​​较​​​​​​​少​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​分​​​​​​​配​​​​​​​更​​​​​​​多​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​。​​​​​​​这​​​​​​​是​​​​​​​一​​​​​​​种​​​​​​​动​​​​​​​态​​​​​​​调​​​​​​​度​​​​​​​算​​​​​​​法​​​​​​​,如​​​​​​​果​​​​​​​请​​​​​​​求​​​​​​​负​​​​​​​载​​​​​​​变​​​​​​​化​​​​​​​较​​​​​​​快​​​​​​​,这​​​​​​​是​​​​​​​一​​​​​​​个​​​​​​​更​​​​​​​好​​​​​​​的​​​​​​​选​​​​​​​择​​​​​​​。​​​​​​​它​​​​​​​最​​​​​​​适​​​​​​​合​​​​​​​于​​​​​​​每​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​都​​​​​​​具​​​​​​​备​​​​​​​基​​​​​​​本​​​​​​​相​​​​​​​同​​​​​​​的​​​​​​​能​​​​​​​力​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​池​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​能​​​​​​​力​​​​​​​不​​​​​​​一​​​​​​​样​​​​​​​,权​​​​​​​重​​​​​​​最​​​​​​​小​​​​​​​连​​​​​​​接​​​​​​​调​​​​​​​度​​​​​​​(weighted least-connection scheduling)就​​​​​​​是​​​​​​​更​​​​​​​好​​​​​​​的​​​​​​​选​​​​​​​择​​​​​​​。​​​​​​​
  • 权​​​​​​​重​​​​​​​最​​​​​​​小​​​​​​​连​​​​​​​接​​​​​​​调​​​​​​​度​​​​​​​(Weighted Least-Connections)— 这​​​​​​​是​​​​​​​缺​​​​​​​省​​​​​​​的​​​​​​​算​​​​​​​法​​​​​​​。​​​​​​​它​​​​​​​将​​​​​​​更​​​​​​​多​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​分​​​​​​​配​​​​​​​给​​​​​​​相​​​​​​​对​​​​​​​其​​​​​​​能​​​​​​​力​​​​​​​具​​​​​​​有​​​​​​​较​​​​​​​少​​​​​​​连​​​​​​​接​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​能​​​​​​​力​​​​​​​以​​​​​​​用​​​​​​​户​​​​​​​分​​​​​​​配​​​​​​​的​​​​​​​权​​​​​​​重​​​​​​​来​​​​​​​表​​​​​​​示​​​​​​​,并​​​​​​​根​​​​​​​据​​​​​​​动​​​​​​​态​​​​​​​负​​​​​​​载​​​​​​​信​​​​​​​息​​​​​​​上​​​​​​​下​​​​​​​调​​​​​​​整​​​​​​​。​​​​​​​当​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​池​​​​​​​包​​​​​​​含​​​​​​​具​​​​​​​有​​​​​​​不​​​​​​​同​​​​​​​能​​​​​​​力​​​​​​​的​​​​​​​硬​​​​​​​件​​​​​​​时​​​​​​​,附​​​​​​​加​​​​​​​的​​​​​​​权​​​​​​​重​​​​​​​使​​​​​​​其​​​​​​​成​​​​​​​为​​​​​​​理​​​​​​​想​​​​​​​的​​​​​​​算​​​​​​​法​​​​​​​。​​​​​​​
  • 基​​​​​​​于​​​​​​​局​​​​​​​部​​​​​​​性​​​​​​​的​​​​​​​最​​​​​​​少​​​​​​​连​​​​​​​接​​​​​​​调​​​​​​​度​​​​​​​(Locality-Based Least-Connection Scheduling)— 将​​​​​​​更​​​​​​​多​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​分​​​​​​​配​​​​​​​给​​​​​​​相​​​​​​​对​​​​​​​其​​​​​​​目​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​具​​​​​​​有​​​​​​​较​​​​​​​少​​​​​​​连​​​​​​​接​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​算​​​​​​​法​​​​​​​用​​​​​​​于​​​​​​​代​​​​​​​理​​​​​​​-缓​​​​​​​存​​​​​​​(proxy-cache)的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​群​​​​​​​集​​​​​​​。​​​​​​​它​​​​​​​根​​​​​​​据​​​​​​​目​​​​​​​标​​​​​​​ IP 地​​​​​​​址​​​​​​​将​​​​​​​数​​​​​​​据​​​​​​​报​​​​​​​文​​​​​​​指​​​​​​​引​​​​​​​到​​​​​​​这​​​​​​​个​​​​​​​地​​​​​​​址​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​,除​​​​​​​非​​​​​​​该​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​超​​​​​​​载​​​​​​​并​​​​​​​有​​​​​​​一​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​处​​​​​​​于​​​​​​​一​​​​​​​半​​​​​​​的​​​​​​​工​​​​​​​作​​​​​​​负​​​​​​​荷​​​​​​​,此​​​​​​​时​​​​​​​它​​​​​​​将​​​​​​​分​​​​​​​配​​​​​​​给​​​​​​​负​​​​​​​荷​​​​​​​最​​​​​​​小​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​
  • 带​​​​​​​复​​​​​​​制​​​​​​​的​​​​​​​基​​​​​​​于​​​​​​​局​​​​​​​部​​​​​​​性​​​​​​​最​​​​​​​小​​​​​​​连​​​​​​​接​​​​​​​调​​​​​​​度​​​​​​​(Locality-Based Least-Connection Scheduling with Replication Scheduling)— 将​​​​​​​更​​​​​​​多​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​分​​​​​​​配​​​​​​​给​​​​​​​相​​​​​​​对​​​​​​​其​​​​​​​目​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​具​​​​​​​有​​​​​​​较​​​​​​​少​​​​​​​连​​​​​​​接​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​算​​​​​​​法​​​​​​​也​​​​​​​用​​​​​​​于​​​​​​​代​​​​​​​理​​​​​​​-缓​​​​​​​存​​​​​​​(proxy-cache)的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​群​​​​​​​集​​​​​​​。​​​​​​​它​​​​​​​和​​​​​​​基​​​​​​​于​​​​​​​局​​​​​​​部​​​​​​​性​​​​​​​最​​​​​​​小​​​​​​​连​​​​​​​接​​​​​​​调​​​​​​​度​​​​​​​的​​​​​​​区​​​​​​​别​​​​​​​是​​​​​​​:它​​​​​​​将​​​​​​​目​​​​​​​标​​​​​​​ IP 地​​​​​​​址​​​​​​​映​​​​​​​射​​​​​​​到​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​一​​​​​​​个​​​​​​​子​​​​​​​集​​​​​​​。​​​​​​​请​​​​​​​求​​​​​​​将​​​​​​​被​​​​​​​发​​​​​​​送​​​​​​​到​​​​​​​这​​​​​​​个​​​​​​​子​​​​​​​集​​​​​​​里​​​​​​​具​​​​​​​有​​​​​​​最​​​​​​​低​​​​​​​连​​​​​​​接​​​​​​​数​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​所​​​​​​​有​​​​​​​目​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​的​​​​​​​节​​​​​​​点​​​​​​​都​​​​​​​超​​​​​​​载​​​​​​​,它​​​​​​​将​​​​​​​整​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​池​​​​​​​里​​​​​​​的​​​​​​​具​​​​​​​有​​​​​​​最​​​​​​​少​​​​​​​连​​​​​​​接​​​​​​​数​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​添​​​​​​​加​​​​​​​到​​​​​​​用​​​​​​​于​​​​​​​该​​​​​​​目​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​子​​​​​​​集​​​​​​​里​​​​​​​,这​​​​​​​样​​​​​​​就​​​​​​​复​​​​​​​制​​​​​​​了​​​​​​​用​​​​​​​于​​​​​​​该​​​​​​​地​​​​​​​址​​​​​​​的​​​​​​​一​​​​​​​个​​​​​​​新​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​然​​​​​​​后​​​​​​​负​​​​​​​载​​​​​​​最​​​​​​​重​​​​​​​的​​​​​​​节​​​​​​​点​​​​​​​将​​​​​​​从​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​子​​​​​​​集​​​​​​​里​​​​​​​删​​​​​​​除​​​​​​​来​​​​​​​防​​​​​​​止​​​​​​​过​​​​​​​度​​​​​​​复​​​​​​​制​​​​​​​。​​​​​​​
  • 源​​​​​​​地​​​​​​​址​​​​​​​散​​​​​​​列​​​​​​​调​​​​​​​度​​​​​​​(Source Hash Scheduling)— 通​​​​​​​过​​​​​​​在​​​​​​​静​​​​​​​态​​​​​​​散​​​​​​​列​​​​​​​表​​​​​​​里​​​​​​​查​​​​​​​找​​​​​​​源​​​​​​​ IP 地​​​​​​​址​​​​​​​来​​​​​​​将​​​​​​​请​​​​​​​求​​​​​​​发​​​​​​​送​​​​​​​到​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​池​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​算​​​​​​​法​​​​​​​用​​​​​​​于​​​​​​​带​​​​​​​有​​​​​​​多​​​​​​​重​​​​​​​防​​​​​​​火​​​​​​​墙​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​。​​​​​​​
而​​​​​​​且​​​​​​​,活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​通​​​​​​​过​​​​​​​简​​​​​​​单​​​​​​​的​​​​​​​ send/expect scripts 动​​​​​​​态​​​​​​​地​​​​​​​监​​​​​​​控​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​的​​​​​​​特​​​​​​​定​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​健​​​​​​​康​​​​​​​状​​​​​​​况​​​​​​​。​​​​​​​为​​​​​​​了​​​​​​​帮​​​​​​​助​​​​​​​检​​​​​​​测​​​​​​​需​​​​​​​要​​​​​​​动​​​​​​​态​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​(如​​​​​​​ HTTPS 或​​​​​​​ SSL)的​​​​​​​健​​​​​​​康​​​​​​​状​​​​​​​况​​​​​​​,您​​​​​​​也​​​​​​​可​​​​​​​以​​​​​​​调​​​​​​​用​​​​​​​外​​​​​​​部​​​​​​​的​​​​​​​执​​​​​​​行​​​​​​​程​​​​​​​序​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​发​​​​​​​生​​​​​​​了​​​​​​​故​​​​​​​障​​​​​​​,活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​将​​​​​​​停​​​​​​​止​​​​​​​发​​​​​​​送​​​​​​​任​​​​​​​务​​​​​​​到​​​​​​​这​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​,直​​​​​​​到​​​​​​​它​​​​​​​恢​​​​​​​复​​​​​​​正​​​​​​​常​​​​​​​操​​​​​​​作​​​​​​​。​​​​​​​
备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​扮​​​​​​​演​​​​​​​备​​​​​​​用​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​角​​​​​​​色​​​​​​​。​​​​​​​LVS 路​​​​​​​由​​​​​​​器​​​​​​​通​​​​​​​过​​​​​​​主​​​​​​​要​​​​​​​的​​​​​​​外​​​​​​​部​​​​​​​公​​​​​​​用​​​​​​​接​​​​​​​口​​​​​​​定​​​​​​​期​​​​​​​地​​​​​​​交​​​​​​​换​​​​​​​ heartbeat 信​​​​​​​息​​​​​​​,在​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​的​​​​​​​情​​​​​​​况​​​​​​​下​​​​​​​则​​​​​​​通​​​​​​​过​​​​​​​专​​​​​​​用​​​​​​​接​​​​​​​口​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​在​​​​​​​预​​​​​​​期​​​​​​​的​​​​​​​时​​​​​​​间​​​​​​​间​​​​​​​隔​​​​​​​里​​​​​​​没​​​​​​​有​​​​​​​接​​​​​​​收​​​​​​​到​​​​​​​ heartbeat 信​​​​​​​息​​​​​​​,它​​​​​​​将​​​​​​​初​​​​​​​始​​​​​​​化​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​并​​​​​​​开​​​​​​​始​​​​​​​执​​​​​​​行​​​​​​​活​​​​​​​动​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​任​​​​​​​务​​​​​​​。​​​​​​​在​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​过​​​​​​​程​​​​​​​中​​​​​​​,备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​用​​​​​​​ ARP spoofing 技​​​​​​​术​​​​​​​接​​​​​​​管​​​​​​​失​​​​​​​效​​​​​​​路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​ VIP 地​​​​​​​址​​​​​​​ — 备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​将​​​​​​​自​​​​​​​己​​​​​​​作​​​​​​​为​​​​​​​发​​​​​​​送​​​​​​​到​​​​​​​失​​​​​​​效​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​ IP 数​​​​​​​据​​​​​​​包​​​​​​​的​​​​​​​目​​​​​​​的​​​​​​​地​​​​​​​。​​​​​​​当​​​​​​​失​​​​​​​效​​​​​​​节​​​​​​​点​​​​​​​恢​​​​​​​复​​​​​​​正​​​​​​​常​​​​​​​服​​​​​​​务​​​​​​​时​​​​​​​,备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​将​​​​​​​再​​​​​​​次​​​​​​​行​​​​​​​使​​​​​​​其​​​​​​​备​​​​​​​份​​​​​​​功​​​​​​​能​​​​​​​。​​​​​​​
The simple, two-tier configuration in 图 1.20 “Two-Tier LVS Topology” is suited best for clusters serving data that does not change very frequently — such as static web pages — because the individual real servers do not automatically synchronize data among themselves.

1.8.2. Three-Tier LVS Topology

图 1.21 “Three-Tier LVS Topology” shows a typical three-tier LVS configuration. In the example, the active LVS router routes the requests from the public network (Internet) to the second tier — real servers. Each real server then accesses a shared data source of a Red Hat cluster in the third tier over the private network.
Three-Tier LVS Topology

图 1.21. Three-Tier LVS Topology

这​​​​​​​种​​​​​​​拓​​​​​​​扑​​​​​​​结​​​​​​​构​​​​​​​适​​​​​​​合​​​​​​​于​​​​​​​繁​​​​​​​忙​​​​​​​的​​​​​​​ FTP 服​​​​​​​务​​​​​​​器​​​​​​​,其​​​​​​​可​​​​​​​访​​​​​​​问​​​​​​​的​​​​​​​数​​​​​​​据​​​​​​​存​​​​​​​储​​​​​​​在​​​​​​​中​​​​​​​央​​​​​​​的​​​​​​​、​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​,而​​​​​​​每​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​都​​​​​​​通​​​​​​​过​​​​​​​导​​​​​​​出​​​​​​​的​​​​​​​ NFS 目​​​​​​​录​​​​​​​或​​​​​​​ Samba 共​​​​​​​享​​​​​​​访​​​​​​​问​​​​​​​该​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​我​​​​​​​们​​​​​​​也​​​​​​​推​​​​​​​荐​​​​​​​在​​​​​​​访​​​​​​​问​​​​​​​用​​​​​​​于​​​​​​​事​​​​​​​务​​​​​​​的​​​​​​​中​​​​​​​央​​​​​​​的​​​​​​​、​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​的​​​​​​​数​​​​​​​据​​​​​​​库​​​​​​​的​​​​​​​网​​​​​​​站​​​​​​​上​​​​​​​应​​​​​​​用​​​​​​​这​​​​​​​种​​​​​​​拓​​​​​​​扑​​​​​​​结​​​​​​​构​​​​​​​。​​​​​​​此​​​​​​​外​​​​​​​,在​​​​​​​ Red Hat 群​​​​​​​集​​​​​​​中​​​​​​​使​​​​​​​用​​​​​​​ active-active 配​​​​​​​置​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​配​​​​​​​置​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​来​​​​​​​同​​​​​​​时​​​​​​​行​​​​​​​使​​​​​​​这​​​​​​​个​​​​​​​两​​​​​​​种​​​​​​​功​​​​​​​能​​​​​​​。​​​​​​​

1.8.3. 路​​​​​​​由​​​​​​​方​​​​​​​法​​​​​​​

您​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​ LVS 里​​​​​​​使​​​​​​​用​​​​​​​网​​​​​​​络​​​​​​​地​​​​​​​址​​​​​​​转​​​​​​​换​​​​​​​(Network Address Translation,NAT)或​​​​​​​直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​。​​​​​​​下​​​​​​​面​​​​​​​的​​​​​​​内​​​​​​​容​​​​​​​简​​​​​​​单​​​​​​​地​​​​​​​描​​​​​​​述​​​​​​​了​​​​​​​在​​​​​​​ LVS 里​​​​​​​使​​​​​​​用​​​​​​​ NAT 路​​​​​​​由​​​​​​​和​​​​​​​直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​。​​​​​​​

1.8.3.1. NAT 路​​​​​​​由​​​​​​​

图 1.22 “LVS Implemented with NAT Routing”, illustrates LVS using NAT routing to move requests between the Internet and a private network.
LVS Implemented with NAT Routing

图 1.22. LVS Implemented with NAT Routing

在​​​​​​​这​​​​​​​个​​​​​​​例​​​​​​​子​​​​​​​中​​​​​​​,活​​​​​​​跃​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​装​​​​​​​有​​​​​​​两​​​​​​​块​​​​​​​网​​​​​​​络​​​​​​​接​​​​​​​口​​​​​​​卡​​​​​​​(NIC)。​​​​​​​Internet 使​​​​​​​用​​​​​​​的​​​​​​​ NIC 在​​​​​​​ eth0 上​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​,它​​​​​​​也​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​别​​​​​​​名​​​​​​​为​​​​​​​ eth0:1 的​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​。​​​​​​​专​​​​​​​用​​​​​​​网​​​​​​​络​​​​​​​接​​​​​​​口​​​​​​​的​​​​​​​ NIC 在​​​​​​​ eth1 上​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​以​​​​​​​及​​​​​​​别​​​​​​​名​​​​​​​为​​​​​​​ eth1:1 的​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​。​​​​​​​在​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​发​​​​​​​生​​​​​​​时​​​​​​​,面​​​​​​​向​​​​​​​ Internet 的​​​​​​​虚​​​​​​​拟​​​​​​​接​​​​​​​口​​​​​​​以​​​​​​​及​​​​​​​专​​​​​​​用​​​​​​​虚​​​​​​​拟​​​​​​​接​​​​​​​口​​​​​​​由​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​同​​​​​​​时​​​​​​​接​​​​​​​管​​​​​​​。​​​​​​​专​​​​​​​用​​​​​​​网​​​​​​​络​​​​​​​中​​​​​​​的​​​​​​​所​​​​​​​有​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​都​​​​​​​使​​​​​​​用​​​​​​​用​​​​​​​于​​​​​​​ NAT 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​来​​​​​​​作​​​​​​​为​​​​​​​与​​​​​​​活​​​​​​​跃​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​通​​​​​​​信​​​​​​​的​​​​​​​默​​​​​​​认​​​​​​​路​​​​​​​由​​​​​​​,这​​​​​​​样​​​​​​​它​​​​​​​们​​​​​​​响​​​​​​​应​​​​​​​来​​​​​​​自​​​​​​​ Internet 的​​​​​​​请​​​​​​​求​​​​​​​的​​​​​​​能​​​​​​​力​​​​​​​就​​​​​​​不​​​​​​​会​​​​​​​被​​​​​​​削​​​​​​​弱​​​​​​​。​​​​​​​
In the example, the LVS router's public LVS floating IP address and private NAT floating IP address are aliased to two physical NICs. While it is possible to associate each floating IP address to its physical device on the LVS router nodes, having more than two NICs is not a requirement.
使​​​​​​​用​​​​​​​这​​​​​​​种​​​​​​​拓​​​​​​​扑​​​​​​​结​​​​​​​构​​​​​​​,活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​接​​​​​​​收​​​​​​​请​​​​​​​求​​​​​​​并​​​​​​​将​​​​​​​其​​​​​​​发​​​​​​​送​​​​​​​到​​​​​​​合​​​​​​​适​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​。​​​​​​​让​​​​​​​后​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​处​​​​​​​理​​​​​​​这​​​​​​​个​​​​​​​请​​​​​​​求​​​​​​​并​​​​​​​将​​​​​​​数​​​​​​​据​​​​​​​包​​​​​​​返​​​​​​​回​​​​​​​给​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​。​​​​​​​LVS 路​​​​​​​由​​​​​​​器​​​​​​​使​​​​​​​网​​​​​​​络​​​​​​​地​​​​​​​址​​​​​​​转​​​​​​​换​​​​​​​来​​​​​​​代​​​​​​​替​​​​​​​带​​​​​​​有​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​挂​​​​​​​公​​​​​​​用​​​​​​​ VIP 地​​​​​​​址​​​​​​​的​​​​​​​数​​​​​​​据​​​​​​​包​​​​​​​里​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​因​​​​​​​为​​​​​​​对​​​​​​​客​​​​​​​户​​​​​​​隐​​​​​​​藏​​​​​​​了​​​​​​​事​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​真​​​​​​​正​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​,所​​​​​​​以​​​​​​​这​​​​​​​个​​​​​​​过​​​​​​​程​​​​​​​被​​​​​​​称​​​​​​​作​​​​​​​ IP 伪​​​​​​​装​​​​​​​(masquerading)。​​​​​​​
使​​​​​​​用​​​​​​​ NAT 路​​​​​​​由​​​​​​​,真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​可​​​​​​​以​​​​​​​是​​​​​​​运​​​​​​​行​​​​​​​不​​​​​​​同​​​​​​​操​​​​​​​作​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​任​​​​​​​何​​​​​​​类​​​​​​​型​​​​​​​的​​​​​​​机​​​​​​​器​​​​​​​。​​​​​​​NAT 的​​​​​​​主​​​​​​​要​​​​​​​缺​​​​​​​点​​​​​​​是​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​可​​​​​​​能​​​​​​​成​​​​​​​为​​​​​​​大​​​​​​​型​​​​​​​部​​​​​​​署​​​​​​​中​​​​​​​的​​​​​​​瓶​​​​​​​颈​​​​​​​,因​​​​​​​为​​​​​​​它​​​​​​​既​​​​​​​得​​​​​​​处​​​​​​​理​​​​​​​转​​​​​​​入​​​​​​​又​​​​​​​得​​​​​​​处​​​​​​​理​​​​​​​转​​​​​​​出​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​。​​​​​​​

1.8.3.2. 直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​

和​​​​​​​ NAT 相​​​​​​​比​​​​​​​,直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​可​​​​​​​以​​​​​​​提​​​​​​​供​​​​​​​增​​​​​​​强​​​​​​​的​​​​​​​性​​​​​​​能​​​​​​​。​​​​​​​直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​允​​​​​​​许​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​处​​​​​​​理​​​​​​​和​​​​​​​直​​​​​​​接​​​​​​​发​​​​​​​送​​​​​​​数​​​​​​​据​​​​​​​包​​​​​​​给​​​​​​​请​​​​​​​求​​​​​​​用​​​​​​​户​​​​​​​而​​​​​​​不​​​​​​​是​​​​​​​通​​​​​​​过​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​传​​​​​​​送​​​​​​​转​​​​​​​出​​​​​​​数​​​​​​​据​​​​​​​包​​​​​​​。​​​​​​​让​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​只​​​​​​​处​​​​​​​理​​​​​​​转​​​​​​​入​​​​​​​的​​​​​​​数​​​​​​​据​​​​​​​包​​​​​​​,直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​降​​​​​​​低​​​​​​​了​​​​​​​网​​​​​​​络​​​​​​​性​​​​​​​能​​​​​​​出​​​​​​​现​​​​​​​问​​​​​​​题​​​​​​​的​​​​​​​可​​​​​​​能​​​​​​​性​​​​​​​。​​​​​​​
LVS Implemented with Direct Routing

图 1.23. LVS Implemented with Direct Routing

在​​​​​​​典​​​​​​​型​​​​​​​的​​​​​​​直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​ LVS 配​​​​​​​置​​​​​​​里​​​​​​​,LVS 路​​​​​​​由​​​​​​​器​​​​​​​通​​​​​​​过​​​​​​​虚​​​​​​​拟​​​​​​​ IP(VIP)接​​​​​​​收​​​​​​​转​​​​​​​入​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​请​​​​​​​求​​​​​​​并​​​​​​​使​​​​​​​用​​​​​​​一​​​​​​​个​​​​​​​调​​​​​​​度​​​​​​​算​​​​​​​法​​​​​​​来​​​​​​​将​​​​​​​请​​​​​​​求​​​​​​​发​​​​​​​送​​​​​​​到​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​。​​​​​​​每​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​都​​​​​​​处​​​​​​​理​​​​​​​请​​​​​​​求​​​​​​​,并​​​​​​​将​​​​​​​响​​​​​​​应​​​​​​​直​​​​​​​接​​​​​​​发​​​​​​​送​​​​​​​给​​​​​​​客​​​​​​​户​​​​​​​而​​​​​​​不​​​​​​​通​​​​​​​过​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​。​​​​​​​直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​提​​​​​​​供​​​​​​​了​​​​​​​可​​​​​​​伸​​​​​​​缩​​​​​​​性​​​​​​​,因​​​​​​​为​​​​​​​可​​​​​​​以​​​​​​​添​​​​​​​加​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​而​​​​​​​不​​​​​​​增​​​​​​​加​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​将​​​​​​​转​​​​​​​出​​​​​​​数​​​​​​​据​​​​​​​包​​​​​​​从​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​发​​​​​​​送​​​​​​​给​​​​​​​客​​​​​​​户​​​​​​​的​​​​​​​负​​​​​​​担​​​​​​​,而​​​​​​​这​​​​​​​在​​​​​​​网​​​​​​​络​​​​​​​负​​​​​​​载​​​​​​​很​​​​​​​重​​​​​​​时​​​​​​​可​​​​​​​能​​​​​​​是​​​​​​​瓶​​​​​​​颈​​​​​​​。​​​​​​​
虽​​​​​​​然​​​​​​​在​​​​​​​ LVS 里​​​​​​​使​​​​​​​用​​​​​​​直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​有​​​​​​​很​​​​​​​多​​​​​​​优​​​​​​​点​​​​​​​,但​​​​​​​它​​​​​​​也​​​​​​​有​​​​​​​缺​​​​​​​陷​​​​​​​。​​​​​​​直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​最​​​​​​​常​​​​​​​见​​​​​​​的​​​​​​​问​​​​​​​题​​​​​​​是​​​​​​​和​​​​​​​ 地​​​​​​​址​​​​​​​解​​​​​​​析​​​​​​​协​​​​​​​议​​​​​​​(Address Resolution Protocol)ARP)相​​​​​​​关​​​​​​​的​​​​​​​。​​​​​​​
In typical situations, a client on the Internet sends a request to an IP address. Network routers typically send requests to their destination by relating IP addresses to a machine's MAC address with ARP. ARP requests are broadcast to all connected machines on a network, and the machine with the correct IP/MAC address combination receives the packet. The IP/MAC associations are stored in an ARP cache, which is cleared periodically (usually every 15 minutes) and refilled with IP/MAC associations.
使​​​​​​​用​​​​​​​直​​​​​​​接​​​​​​​路​​​​​​​由​​​​​​​的​​​​​​​ LVS 配​​​​​​​置​​​​​​​里​​​​​​​和​​​​​​​ ARP 请​​​​​​​求​​​​​​​相​​​​​​​关​​​​​​​的​​​​​​​问​​​​​​​题​​​​​​​是​​​​​​​:因​​​​​​​为​​​​​​​到​​​​​​​一​​​​​​​个​​​​​​​ IP 地​​​​​​​址​​​​​​​的​​​​​​​客​​​​​​​户​​​​​​​请​​​​​​​求​​​​​​​必​​​​​​​须​​​​​​​和​​​​​​​一​​​​​​​个​​​​​​​ MAC 地​​​​​​​址​​​​​​​相​​​​​​​关​​​​​​​联​​​​​​​才​​​​​​​能​​​​​​​被​​​​​​​处​​​​​​​理​​​​​​​,LVS 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​虚​​​​​​​拟​​​​​​​ IP 地​​​​​​​址​​​​​​​也​​​​​​​必​​​​​​​须​​​​​​​关​​​​​​​联​​​​​​​到​​​​​​​一​​​​​​​个​​​​​​​ MAC。​​​​​​​然​​​​​​​而​​​​​​​,因​​​​​​​为​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​和​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​具​​​​​​​有​​​​​​​相​​​​​​​同​​​​​​​的​​​​​​​ VIP,ARP 请​​​​​​​求​​​​​​​被​​​​​​​广​​​​​​​播​​​​​​​至​​​​​​​和​​​​​​​这​​​​​​​个​​​​​​​ VIP 相​​​​​​​关​​​​​​​联​​​​​​​的​​​​​​​所​​​​​​​有​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​这​​​​​​​会​​​​​​​导​​​​​​​致​​​​​​​几​​​​​​​个​​​​​​​问​​​​​​​题​​​​​​​,如​​​​​​​直​​​​​​​接​​​​​​​和​​​​​​​某​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​相​​​​​​​关​​​​​​​联​​​​​​​或​​​​​​​直​​​​​​​接​​​​​​​处​​​​​​​理​​​​​​​请​​​​​​​求​​​​​​​的​​​​​​​ VIP,将​​​​​​​完​​​​​​​全​​​​​​​绕​​​​​​​过​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​并​​​​​​​使​​​​​​​ LVS 配​​​​​​​置​​​​​​​不​​​​​​​起​​​​​​​作​​​​​​​用​​​​​​​。​​​​​​​即​​​​​​​使​​​​​​​是​​​​​​​带​​​​​​​有​​​​​​​强​​​​​​​大​​​​​​​的​​​​​​​ CPU、​​​​​​​可​​​​​​​以​​​​​​​快​​​​​​​速​​​​​​​响​​​​​​​应​​​​​​​客​​​​​​​户​​​​​​​请​​​​​​​求​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​也​​​​​​​未​​​​​​​必​​​​​​​能​​​​​​​解​​​​​​​决​​​​​​​这​​​​​​​个​​​​​​​问​​​​​​​题​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​负​​​​​​​载​​​​​​​很​​​​​​​重​​​​​​​,它​​​​​​​对​​​​​​​ ARP 的​​​​​​​响​​​​​​​应​​​​​​​可​​​​​​​能​​​​​​​会​​​​​​​比​​​​​​​较​​​​​​​空​​​​​​​闲​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​更​​​​​​​慢​​​​​​​,后​​​​​​​者​​​​​​​分​​​​​​​配​​​​​​​了​​​​​​​请​​​​​​​求​​​​​​​客​​​​​​​户​​​​​​​的​​​​​​​ ARP 缓​​​​​​​存​​​​​​​里​​​​​​​的​​​​​​​ VIP 且​​​​​​​响​​​​​​​应​​​​​​​更​​​​​​​快​​​​​​​。​​​​​​​
要​​​​​​​解​​​​​​​决​​​​​​​这​​​​​​​个​​​​​​​问​​​​​​​题​​​​​​​,转​​​​​​​入​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​应​​​​​​​该​​​​​​​只​​​​​​​和​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​ VIP 相​​​​​​​关​​​​​​​联​​​​​​​,这​​​​​​​样​​​​​​​就​​​​​​​会​​​​​​​正​​​​​​​确​​​​​​​地​​​​​​​处​​​​​​​理​​​​​​​请​​​​​​​求​​​​​​​并​​​​​​​将​​​​​​​它​​​​​​​们​​​​​​​发​​​​​​​送​​​​​​​到​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​池​​​​​​​。​​​​​​​这​​​​​​​可​​​​​​​以​​​​​​​用​​​​​​​ arptables 数​​​​​​​据​​​​​​​包​​​​​​​过​​​​​​​滤​​​​​​​工​​​​​​​具​​​​​​​来​​​​​​​实​​​​​​​现​​​​​​​。​​​​​​​

1.8.4. 持​​​​​​​久​​​​​​​性​​​​​​​和​​​​​​​防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​

在​​​​​​​某​​​​​​​些​​​​​​​情​​​​​​​况​​​​​​​下​​​​​​​,对​​​​​​​于​​​​​​​客​​​​​​​户​​​​​​​来​​​​​​​说​​​​​​​,重​​​​​​​新​​​​​​​连​​​​​​​接​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​、​​​​​​​而​​​​​​​不​​​​​​​是​​​​​​​用​​​​​​​ LVS 负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​算​​​​​​​法​​​​​​​将​​​​​​​请​​​​​​​求​​​​​​​发​​​​​​​送​​​​​​​到​​​​​​​最​​​​​​​合​​​​​​​适​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​是​​​​​​​可​​​​​​​取​​​​​​​的​​​​​​​。​​​​​​​这​​​​​​​样​​​​​​​的​​​​​​​例​​​​​​​子​​​​​​​包​​​​​​​括​​​​​​​多​​​​​​​屏​​​​​​​幕​​​​​​​的​​​​​​​ web 表​​​​​​​单​​​​​​​、​​​​​​​cookies、​​​​​​​SSL 和​​​​​​​ FTP 连​​​​​​​接​​​​​​​。​​​​​​​在​​​​​​​这​​​​​​​些​​​​​​​情​​​​​​​况​​​​​​​下​​​​​​​,除​​​​​​​非​​​​​​​保​​​​​​​留​​​​​​​了​​​​​​​上​​​​​​​下​​​​​​​文​​​​​​​的​​​​​​​相​​​​​​​同​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​来​​​​​​​处​​​​​​​理​​​​​​​事​​​​​​​务​​​​​​​用​​​​​​​,否​​​​​​​则​​​​​​​客​​​​​​​户​​​​​​​端​​​​​​​可​​​​​​​能​​​​​​​不​​​​​​​能​​​​​​​正​​​​​​​确​​​​​​​工​​​​​​​作​​​​​​​。​​​​​​​LVS 提​​​​​​​供​​​​​​​两​​​​​​​个​​​​​​​不​​​​​​​同​​​​​​​的​​​​​​​功​​​​​​​能​​​​​​​来​​​​​​​处​​​​​​​理​​​​​​​这​​​​​​​种​​​​​​​情​​​​​​​况​​​​​​​:持​​​​​​​久​​​​​​​性​​​​​​​(persistence)和​​​​​​​防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​(firewall mark)。​​​​​​​

1.8.4.1. Persistence

启​​​​​​​用​​​​​​​后​​​​​​​,持​​​​​​​久​​​​​​​性​​​​​​​起​​​​​​​到​​​​​​​定​​​​​​​时​​​​​​​器​​​​​​​的​​​​​​​作​​​​​​​用​​​​​​​。​​​​​​​当​​​​​​​客​​​​​​​户​​​​​​​连​​​​​​​接​​​​​​​一​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​时​​​​​​​,LVS 会​​​​​​​在​​​​​​​指​​​​​​​定​​​​​​​的​​​​​​​时​​​​​​​间​​​​​​​内​​​​​​​记​​​​​​​住​​​​​​​最​​​​​​​后​​​​​​​的​​​​​​​连​​​​​​​接​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​具​​​​​​​有​​​​​​​相​​​​​​​同​​​​​​​ IP 地​​​​​​​址​​​​​​​的​​​​​​​客​​​​​​​户​​​​​​​在​​​​​​​这​​​​​​​段​​​​​​​时​​​​​​​间​​​​​​​内​​​​​​​再​​​​​​​次​​​​​​​进​​​​​​​行​​​​​​​连​​​​​​​接​​​​​​​,它​​​​​​​将​​​​​​​被​​​​​​​送​​​​​​​往​​​​​​​和​​​​​​​上​​​​​​​次​​​​​​​连​​​​​​​接​​​​​​​相​​​​​​​同​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​ — 忽​​​​​​​略​​​​​​​负​​​​​​​载​​​​​​​平​​​​​​​衡​​​​​​​机​​​​​​​制​​​​​​​。​​​​​​​而​​​​​​​当​​​​​​​连​​​​​​​接​​​​​​​在​​​​​​​这​​​​​​​段​​​​​​​时​​​​​​​间​​​​​​​外​​​​​​​发​​​​​​​生​​​​​​​,它​​​​​​​会​​​​​​​按​​​​​​​照​​​​​​​适​​​​​​​当​​​​​​​的​​​​​​​调​​​​​​​度​​​​​​​规​​​​​​​则​​​​​​​进​​​​​​​行​​​​​​​处​​​​​​​理​​​​​​​。​​​​​​​
持​​​​​​​久​​​​​​​性​​​​​​​也​​​​​​​允​​​​​​​许​​​​​​​您​​​​​​​指​​​​​​​定​​​​​​​适​​​​​​​用​​​​​​​于​​​​​​​客​​​​​​​户​​​​​​​ IP 地​​​​​​​址​​​​​​​的​​​​​​​子​​​​​​​网​​​​​​​掩​​​​​​​码​​​​​​​,它​​​​​​​将​​​​​​​连​​​​​​​接​​​​​​​按​​​​​​​子​​​​​​​网​​​​​​​分​​​​​​​组​​​​​​​,控​​​​​​​制​​​​​​​哪​​​​​​​些​​​​​​​地​​​​​​​址​​​​​​​具​​​​​​​有​​​​​​​更​​​​​​​高​​​​​​​的​​​​​​​持​​​​​​​久​​​​​​​性​​​​​​​级​​​​​​​别​​​​​​​。​​​​​​​
对​​​​​​​于​​​​​​​使​​​​​​​用​​​​​​​多​​​​​​​个​​​​​​​端​​​​​​​口​​​​​​​通​​​​​​​信​​​​​​​的​​​​​​​协​​​​​​​议​​​​​​​(如​​​​​​​ FTP)来​​​​​​​说​​​​​​​,根​​​​​​​据​​​​​​​端​​​​​​​口​​​​​​​进​​​​​​​行​​​​​​​连​​​​​​​接​​​​​​​分​​​​​​​组​​​​​​​是​​​​​​​很​​​​​​​重​​​​​​​要​​​​​​​的​​​​​​​。​​​​​​​然​​​​​​​而​​​​​​​,持​​​​​​​久​​​​​​​性​​​​​​​不​​​​​​​是​​​​​​​处​​​​​​​理​​​​​​​根​​​​​​​据​​​​​​​端​​​​​​​口​​​​​​​进​​​​​​​行​​​​​​​连​​​​​​​接​​​​​​​分​​​​​​​组​​​​​​​的​​​​​​​效​​​​​​​率​​​​​​​最​​​​​​​高​​​​​​​的​​​​​​​办​​​​​​​法​​​​​​​。​​​​​​​对​​​​​​​于​​​​​​​这​​​​​​​样​​​​​​​的​​​​​​​情​​​​​​​况​​​​​​​,您​​​​​​​最​​​​​​​好​​​​​​​使​​​​​​​用​​​​​​​防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​(firewall mark)。​​​​​​​

1.8.4.2. 防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​

根​​​​​​​据​​​​​​​协​​​​​​​议​​​​​​​或​​​​​​​相​​​​​​​关​​​​​​​协​​​​​​​议​​​​​​​组​​​​​​​使​​​​​​​用​​​​​​​不​​​​​​​同​​​​​​​的​​​​​​​端​​​​​​​口​​​​​​​,防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​是​​​​​​​容​​​​​​​易​​​​​​​且​​​​​​​高​​​​​​​效​​​​​​​率​​​​​​​的​​​​​​​方​​​​​​​法​​​​​​​。​​​​​​​例​​​​​​​如​​​​​​​,如​​​​​​​果​​​​​​​ LVS 部​​​​​​​署​​​​​​​在​​​​​​​一​​​​​​​个​​​​​​​电​​​​​​​子​​​​​​​商​​​​​​​务​​​​​​​网​​​​​​​站​​​​​​​里​​​​​​​,防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​可​​​​​​​以​​​​​​​将​​​​​​​ HTTP 连​​​​​​​接​​​​​​​绑​​​​​​​定​​​​​​​在​​​​​​​端​​​​​​​口​​​​​​​ 80、​​​​​​​安​​​​​​​全​​​​​​​ HTTP 连​​​​​​​接​​​​​​​绑​​​​​​​定​​​​​​​在​​​​​​​端​​​​​​​口​​​​​​​ 443。​​​​​​​通​​​​​​​过​​​​​​​根​​​​​​​据​​​​​​​每​​​​​​​个​​​​​​​协​​​​​​​议​​​​​​​为​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​分​​​​​​​配​​​​​​​相​​​​​​​同​​​​​​​的​​​​​​​防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​,事​​​​​​​务​​​​​​​的​​​​​​​状​​​​​​​态​​​​​​​信​​​​​​​息​​​​​​​可​​​​​​​以​​​​​​​保​​​​​​​留​​​​​​​,因​​​​​​​为​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​在​​​​​​​连​​​​​​​接​​​​​​​打​​​​​​​开​​​​​​​后​​​​​​​会​​​​​​​将​​​​​​​所​​​​​​​有​​​​​​​的​​​​​​​请​​​​​​​求​​​​​​​转​​​​​​​发​​​​​​​到​​​​​​​相​​​​​​​同​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​
因​​​​​​​为​​​​​​​它​​​​​​​效​​​​​​​率​​​​​​​高​​​​​​​且​​​​​​​易​​​​​​​于​​​​​​​使​​​​​​​用​​​​​​​,对​​​​​​​于​​​​​​​连​​​​​​​接​​​​​​​分​​​​​​​组​​​​​​​,LVS 管​​​​​​​理​​​​​​​员​​​​​​​应​​​​​​​该​​​​​​​尽​​​​​​​可​​​​​​​能​​​​​​​用​​​​​​​防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​代​​​​​​​替​​​​​​​持​​​​​​​久​​​​​​​性​​​​​​​。​​​​​​​然​​​​​​​而​​​​​​​,您​​​​​​​应​​​​​​​该​​​​​​​仍​​​​​​​然​​​​​​​将​​​​​​​持​​​​​​​久​​​​​​​性​​​​​​​和​​​​​​​防​​​​​​​火​​​​​​​墙​​​​​​​标​​​​​​​记​​​​​​​一​​​​​​​起​​​​​​​加​​​​​​​入​​​​​​​到​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​里​​​​​​​,这​​​​​​​样​​​​​​​才​​​​​​​能​​​​​​​确​​​​​​​保​​​​​​​客​​​​​​​户​​​​​​​在​​​​​​​足​​​​​​​够​​​​​​​长​​​​​​​的​​​​​​​时​​​​​​​间​​​​​​​内​​​​​​​重​​​​​​​新​​​​​​​连​​​​​​​接​​​​​​​到​​​​​​​相​​​​​​​同​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​

1.9. 群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​

Red Hat Cluster Suite 提​​​​​​​供​​​​​​​不​​​​​​​同​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​来​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​您​​​​​​​的​​​​​​​ Red Hat 群​​​​​​​集​​​​​​​。​​​​​​​本​​​​​​​节​​​​​​​是​​​​​​​对​​​​​​​ Red Hat Cluster Suite 里​​​​​​​可​​​​​​​用​​​​​​​的​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​的​​​​​​​概​​​​​​​述​​​​​​​:

1.9.1. Conga

Conga 是​​​​​​​为​​​​​​​ Red Hat 群​​​​​​​集​​​​​​​和​​​​​​​存​​​​​​​储​​​​​​​提​​​​​​​供​​​​​​​集​​​​​​​中​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​的​​​​​​​软​​​​​​​件​​​​​​​组​​​​​​​件​​​​​​​集​​​​​​​合​​​​​​​。​​​​​​​Conga 具​​​​​​​有​​​​​​​下​​​​​​​面​​​​​​​的​​​​​​​主​​​​​​​要​​​​​​​特​​​​​​​征​​​​​​​:
  • 管​​​​​​​理​​​​​​​群​​​​​​​集​​​​​​​和​​​​​​​存​​​​​​​储​​​​​​​的​​​​​​​ web 界​​​​​​​面​​​​​​​
  • 自​​​​​​​动​​​​​​​部​​​​​​​署​​​​​​​群​​​​​​​集​​​​​​​数​​​​​​​据​​​​​​​和​​​​​​​支​​​​​​​持​​​​​​​软​​​​​​​件​​​​​​​包​​​​​​​
  • 易​​​​​​​于​​​​​​​集​​​​​​​成​​​​​​​现​​​​​​​有​​​​​​​群​​​​​​​集​​​​​​​
  • 不​​​​​​​需​​​​​​​要​​​​​​​重​​​​​​​验​​​​​​​证​​​​​​​
  • 集​​​​​​​成​​​​​​​群​​​​​​​集​​​​​​​状​​​​​​​态​​​​​​​和​​​​​​​日​​​​​​​志​​​​​​​
  • 用​​​​​​​户​​​​​​​权​​​​​​​限​​​​​​​的​​​​​​​细​​​​​​​颗​​​​​​​粒​​​​​​​度​​​​​​​控​​​​​​​制​​​​​​​
Conga 里​​​​​​​的​​​​​​​主​​​​​​​要​​​​​​​组​​​​​​​件​​​​​​​是​​​​​​​ luci 和​​​​​​​ ricci,它​​​​​​​们​​​​​​​都​​​​​​​是​​​​​​​可​​​​​​​单​​​​​​​独​​​​​​​安​​​​​​​装​​​​​​​的​​​​​​​。​​​​​​​luci 是​​​​​​​一​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​组​​​​​​​件​​​​​​​,它​​​​​​​运​​​​​​​行​​​​​​​在​​​​​​​一​​​​​​​台​​​​​​​机​​​​​​​器​​​​​​​上​​​​​​​并​​​​​​​通​​​​​​​过​​​​​​​ ricci 与​​​​​​​多​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​和​​​​​​​机​​​​​​​器​​​​​​​通​​​​​​​信​​​​​​​。​​​​​​​ricci 是​​​​​​​运​​​​​​​行​​​​​​​在​​​​​​​ Conga 管​​​​​​​理​​​​​​​的​​​​​​​每​​​​​​​台​​​​​​​机​​​​​​​器​​​​​​​(群​​​​​​​集​​​​​​​成​​​​​​​员​​​​​​​或​​​​​​​独​​​​​​​立​​​​​​​的​​​​​​​机​​​​​​​器​​​​​​​)上​​​​​​​的​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
luci 可​​​​​​​以​​​​​​​通​​​​​​​过​​​​​​​ Web 浏​​​​​​​览​​​​​​​器​​​​​​​访​​​​​​​问​​​​​​​,它​​​​​​​提​​​​​​​供​​​​​​​ 3 个​​​​​​​可​​​​​​​以​​​​​​​从​​​​​​​下​​​​​​​面​​​​​​​的​​​​​​​标​​​​​​​签​​​​​​​页​​​​​​​访​​​​​​​问​​​​​​​的​​​​​​​主​​​​​​​要​​​​​​​功​​​​​​​能​​​​​​​:
  • homebase — 提​​​​​​​供​​​​​​​添​​​​​​​加​​​​​​​和​​​​​​​删​​​​​​​除​​​​​​​机​​​​​​​器​​​​​​​、​​​​​​​添​​​​​​​加​​​​​​​和​​​​​​​删​​​​​​​除​​​​​​​用​​​​​​​户​​​​​​​和​​​​​​​配​​​​​​​置​​​​​​​用​​​​​​​户​​​​​​​权​​​​​​​限​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​只​​​​​​​有​​​​​​​系​​​​​​​统​​​​​​​管​​​​​​​理​​​​​​​员​​​​​​​才​​​​​​​被​​​​​​​允​​​​​​​许​​​​​​​访​​​​​​​问​​​​​​​此​​​​​​​标​​​​​​​签​​​​​​​页​​​​​​​。​​​​​​​
  • cluster — 提​​​​​​​供​​​​​​​创​​​​​​​建​​​​​​​和​​​​​​​配​​​​​​​置​​​​​​​群​​​​​​​集​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​luci 的​​​​​​​每​​​​​​​个​​​​​​​实​​​​​​​例​​​​​​​都​​​​​​​列​​​​​​​出​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​ luci 设​​​​​​​立​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​。​​​​​​​系​​​​​​​统​​​​​​​管​​​​​​​理​​​​​​​员​​​​​​​可​​​​​​​以​​​​​​​管​​​​​​​理​​​​​​​此​​​​​​​标​​​​​​​签​​​​​​​页​​​​​​​上​​​​​​​列​​​​​​​出​​​​​​​的​​​​​​​所​​​​​​​有​​​​​​​群​​​​​​​集​​​​​​​。​​​​​​​其​​​​​​​他​​​​​​​用​​​​​​​户​​​​​​​只​​​​​​​能​​​​​​​管​​​​​​​理​​​​​​​其​​​​​​​具​​​​​​​有​​​​​​​管​​​​​​​理​​​​​​​权​​​​​​​限​​​​​​​(由​​​​​​​管​​​​​​​理​​​​​​​员​​​​​​​赋​​​​​​​予​​​​​​​)的​​​​​​​群​​​​​​​集​​​​​​​。​​​​​​​
  • storage — 提​​​​​​​供​​​​​​​远​​​​​​​程​​​​​​​管​​​​​​​理​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​您​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​属​​​​​​​于​​​​​​​或​​​​​​​不​​​​​​​属​​​​​​​于​​​​​​​群​​​​​​​集​​​​​​​的​​​​​​​机​​​​​​​器​​​​​​​上​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​标​​​​​​​签​​​​​​​页​​​​​​​上​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​
要​​​​​​​管​​​​​​​理​​​​​​​群​​​​​​​集​​​​​​​或​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​,管​​​​​​​理​​​​​​​员​​​​​​​要​​​​​​​将​​​​​​​群​​​​​​​集​​​​​​​或​​​​​​​机​​​​​​​器​​​​​​​添​​​​​​​加​​​​​​​(或​​​​​​​注​​​​​​​册​​​​​​​)到​​​​​​​ luci 服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​当​​​​​​​群​​​​​​​集​​​​​​​或​​​​​​​机​​​​​​​器​​​​​​​注​​​​​​​册​​​​​​​了​​​​​​​ luci,每​​​​​​​台​​​​​​​机​​​​​​​器​​​​​​​的​​​​​​​全​​​​​​​限​​​​​​​定​​​​​​​主​​​​​​​机​​​​​​​名​​​​​​​或​​​​​​​ IP 地​​​​​​​址​​​​​​​都​​​​​​​保​​​​​​​存​​​​​​​在​​​​​​​ luci 数​​​​​​​据​​​​​​​库​​​​​​​里​​​​​​​。​​​​​​​
您​​​​​​​可​​​​​​​以​​​​​​​将​​​​​​​一​​​​​​​个​​​​​​​ luci 实​​​​​​​例​​​​​​​的​​​​​​​数​​​​​​​据​​​​​​​库​​​​​​​传​​​​​​​递​​​​​​​给​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​ luci 实​​​​​​​例​​​​​​​。​​​​​​​这​​​​​​​种​​​​​​​能​​​​​​​力​​​​​​​提​​​​​​​供​​​​​​​了​​​​​​​复​​​​​​​制​​​​​​​ luci 服​​​​​​​务​​​​​​​器​​​​​​​实​​​​​​​例​​​​​​​的​​​​​​​方​​​​​​​法​​​​​​​以​​​​​​​及​​​​​​​高​​​​​​​效​​​​​​​的​​​​​​​升​​​​​​​级​​​​​​​和​​​​​​​测​​​​​​​试​​​​​​​的​​​​​​​途​​​​​​​径​​​​​​​。​​​​​​​当​​​​​​​ luci 实​​​​​​​例​​​​​​​刚​​​​​​​被​​​​​​​安​​​​​​​装​​​​​​​时​​​​​​​,它​​​​​​​的​​​​​​​数​​​​​​​据​​​​​​​库​​​​​​​是​​​​​​​空​​​​​​​的​​​​​​​。​​​​​​​然​​​​​​​而​​​​​​​,当​​​​​​​部​​​​​​​署​​​​​​​新​​​​​​​的​​​​​​​ luci 服​​​​​​​务​​​​​​​器​​​​​​​时​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​从​​​​​​​现​​​​​​​有​​​​​​​的​​​​​​​ luci 服​​​​​​​务​​​​​​​器​​​​​​​导​​​​​​​入​​​​​​​部​​​​​​​分​​​​​​​或​​​​​​​全​​​​​​​部​​​​​​​的​​​​​​​ luci 数​​​​​​​据​​​​​​​库​​​​​​​。​​​​​​​
在​​​​​​​初​​​​​​​始​​​​​​​安​​​​​​​装​​​​​​​时​​​​​​​,每​​​​​​​个​​​​​​​ luci 实​​​​​​​例​​​​​​​都​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​用​​​​​​​户​​​​​​​ — admin。​​​​​​​只​​​​​​​有​​​​​​​ admin 可​​​​​​​以​​​​​​​将​​​​​​​系​​​​​​​统​​​​​​​添​​​​​​​加​​​​​​​到​​​​​​​ luci 服​​​​​​​务​​​​​​​器​​​​​​​里​​​​​​​。​​​​​​​而​​​​​​​且​​​​​​​,admin 用​​​​​​​户​​​​​​​可​​​​​​​以​​​​​​​创​​​​​​​建​​​​​​​其​​​​​​​他​​​​​​​的​​​​​​​用​​​​​​​户​​​​​​​帐​​​​​​​号​​​​​​​并​​​​​​​决​​​​​​​定​​​​​​​哪​​​​​​​些​​​​​​​用​​​​​​​户​​​​​​​可​​​​​​​以​​​​​​​访​​​​​​​问​​​​​​​在​​​​​​​ luci 里​​​​​​​注​​​​​​​册​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​和​​​​​​​机​​​​​​​器​​​​​​​。​​​​​​​您​​​​​​​也​​​​​​​可​​​​​​​以​​​​​​​将​​​​​​​用​​​​​​​户​​​​​​​批​​​​​​​量​​​​​​​地​​​​​​​导​​​​​​​入​​​​​​​到​​​​​​​新​​​​​​​的​​​​​​​ luci 服​​​​​​​务​​​​​​​器​​​​​​​里​​​​​​​,就​​​​​​​如​​​​​​​同​​​​​​​导​​​​​​​入​​​​​​​群​​​​​​​集​​​​​​​和​​​​​​​节​​​​​​​点​​​​​​​一​​​​​​​样​​​​​​​。​​​​​​​
当​​​​​​​机​​​​​​​器​​​​​​​被​​​​​​​加​​​​​​​入​​​​​​​ luci 服​​​​​​​务​​​​​​​器​​​​​​​进​​​​​​​行​​​​​​​管​​​​​​​理​​​​​​​时​​​​​​​,将​​​​​​​进​​​​​​​行​​​​​​​一​​​​​​​次​​​​​​​验​​​​​​​证​​​​​​​。​​​​​​​此​​​​​​​后​​​​​​​不​​​​​​​需​​​​​​​要​​​​​​​再​​​​​​​进​​​​​​​行​​​​​​​验​​​​​​​证​​​​​​​(除​​​​​​​非​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​证​​​​​​​书​​​​​​​被​​​​​​​ CA 调​​​​​​​用​​​​​​​)。​​​​​​​然​​​​​​​后​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​通​​​​​​​过​​​​​​​ luci 用​​​​​​​户​​​​​​​界​​​​​​​面​​​​​​​远​​​​​​​程​​​​​​​地​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​群​​​​​​​集​​​​​​​和​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​。​​​​​​​luci 和​​​​​​​ ricci 通​​​​​​​过​​​​​​​ XML 进​​​​​​​行​​​​​​​相​​​​​​​互​​​​​​​通​​​​​​​信​​​​​​​。​​​​​​​
下​​​​​​​面​​​​​​​的​​​​​​​图​​​​​​​例​​​​​​​展​​​​​​​示​​​​​​​了​​​​​​​ 3 个​​​​​​​主​​​​​​​要​​​​​​​的​​​​​​​ luci 标​​​​​​​签​​​​​​​页​​​​​​​:homebase、​​​​​​​cluster 和​​​​​​​ storage。​​​​​​​
关​​​​​​​于​​​​​​​ Conga 的​​​​​​​更​​​​​​​多​​​​​​​信​​​​​​​息​​​​​​​,请​​​​​​​参​​​​​​​考​​​​​​​《​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​红​​​​​​​帽​​​​​​​群​​​​​​​集​​​​​​​》​​​​​​​以​​​​​​​及​​​​​​​ luci 服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​在​​​​​​​线​​​​​​​帮​​​​​​​助​​​​​​​。​​​​​​​
luci 的​​​​​​​ homebase 标​​​​​​​签​​​​​​​页​​​​​​​

图 1.24. luci 的​​​​​​​ homebase 标​​​​​​​签​​​​​​​页​​​​​​​

luci 的​​​​​​​ cluster 标​​​​​​​签​​​​​​​页​​​​​​​

图 1.25. luci 的​​​​​​​ cluster 标​​​​​​​签​​​​​​​页​​​​​​​

luci 的​​​​​​​ storage 标​​​​​​​签​​​​​​​页​​​​​​​

图 1.26. luci 的​​​​​​​ storage 标​​​​​​​签​​​​​​​页​​​​​​​

1.9.2. 群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​图​​​​​​​形​​​​​​​界​​​​​​​面​​​​​​​

This section provides an overview of the system-config-cluster cluster administration graphical user interface (GUI) available with Red Hat Cluster Suite. The GUI is for use with the cluster infrastructure and the high-availability service management components (refer to 第 1.3 节 “Cluster Infrastructure” and 第 1.4 节 “高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​”). The GUI consists of two major functions: the Cluster Configuration Tool and the Cluster Status Tool. The Cluster Configuration Tool provides the capability to create, edit, and propagate the cluster configuration file (/etc/cluster/cluster.conf). The Cluster Status Tool provides the capability to manage high-availability services. The following sections summarize those functions.

1.9.2.1. Cluster Configuration Tool

You can access the Cluster Configuration Tool (图 1.27 “Cluster Configuration Tool) through the Cluster Configuration tab in the Cluster Administration GUI.
Cluster Configuration Tool

图 1.27. Cluster Configuration Tool

Cluster Configuration Tool 以​​​​​​​分​​​​​​​层​​​​​​​结​​​​​​​构​​​​​​​的​​​​​​​形​​​​​​​式​​​​​​​在​​​​​​​面​​​​​​​板​​​​​​​的​​​​​​​左​​​​​​​边​​​​​​​显​​​​​​​示​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​(/etc/cluster/cluster.conf)里​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​组​​​​​​​件​​​​​​​。​​​​​​​组​​​​​​​件​​​​​​​命​​​​​​​左​​​​​​​边​​​​​​​的​​​​​​​三​​​​​​​角​​​​​​​形​​​​​​​图​​​​​​​标​​​​​​​表​​​​​​​示​​​​​​​这​​​​​​​个​​​​​​​组​​​​​​​件​​​​​​​分​​​​​​​配​​​​​​​有​​​​​​​一​​​​​​​个​​​​​​​或​​​​​​​多​​​​​​​个​​​​​​​下​​​​​​​级​​​​​​​组​​​​​​​件​​​​​​​。​​​​​​​点​​​​​​​击​​​​​​​这​​​​​​​个​​​​​​​三​​​​​​​角​​​​​​​图​​​​​​​标​​​​​​​可​​​​​​​以​​​​​​​展​​​​​​​开​​​​​​​组​​​​​​​件​​​​​​​下​​​​​​​的​​​​​​​目​​​​​​​录​​​​​​​树​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​图​​​​​​​形​​​​​​​界​​​​​​​面​​​​​​​显​​​​​​​示​​​​​​​如​​​​​​​下​​​​​​​的​​​​​​​组​​​​​​​件​​​​​​​:
  • 「​​​​​​​Cluster Nodes」​​​​​​​ — 显​​​​​​​示​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​。​​​​​​​节​​​​​​​点​​​​​​​名​​​​​​​出​​​​​​​现​​​​​​​在​​​​​​​「​​​​​​​Cluster Nodes」​​​​​​​的​​​​​​​下​​​​​​​级​​​​​​​列​​​​​​​表​​​​​​​里​​​​​​​。​​​​​​​使​​​​​​​用​​​​​​​右​​​​​​​栏​​​​​​​底​​​​​​​部​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​按​​​​​​​钮​​​​​​​(在​​​​​​​「​​​​​​​Properties」​​​​​​​下​​​​​​​面​​​​​​​),您​​​​​​​可​​​​​​​以​​​​​​​添​​​​​​​加​​​​​​​节​​​​​​​点​​​​​​​、​​​​​​​删​​​​​​​除​​​​​​​节​​​​​​​点​​​​​​​、​​​​​​​编​​​​​​​辑​​​​​​​节​​​​​​​点​​​​​​​属​​​​​​​性​​​​​​​以​​​​​​​及​​​​​​​为​​​​​​​每​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​配​​​​​​​置​​​​​​​ fencing 方​​​​​​​法​​​​​​​。​​​​​​​
  • 「​​​​​​​Fence Devices」​​​​​​​ — 显​​​​​​​示​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​。​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​出​​​​​​​现​​​​​​​在​​​​​​​「​​​​​​​Fence Devices」​​​​​​​的​​​​​​​下​​​​​​​级​​​​​​​列​​​​​​​表​​​​​​​里​​​​​​​。​​​​​​​使​​​​​​​用​​​​​​​右​​​​​​​栏​​​​​​​底​​​​​​​部​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​按​​​​​​​钮​​​​​​​(在​​​​​​​「​​​​​​​Properties」​​​​​​​下​​​​​​​面​​​​​​​),您​​​​​​​可​​​​​​​以​​​​​​​添​​​​​​​加​​​​​​​、​​​​​​​删​​​​​​​除​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​、​​​​​​​编​​​​​​​辑​​​​​​​其​​​​​​​属​​​​​​​性​​​​​​​。​​​​​​​在​​​​​​​配​​​​​​​置​​​​​​​保​​​​​​​护​​​​​​​之​​​​​​​前​​​​​​​,您​​​​​​​必​​​​​​​须​​​​​​​先​​​​​​​为​​​​​​​每​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​定​​​​​​​义​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​(用​​​​​​​「​​​​​​​Manage Fencing For This Node」​​​​​​​按​​​​​​​钮​​​​​​​)。​​​​​​​
  • 「​​​​​​​Managed Resources」​​​​​​​ — 显​​​​​​​示​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​、​​​​​​​资​​​​​​​源​​​​​​​和​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​
    • 「​​​​​​​Failover Domains」​​​​​​​ — 配​​​​​​​置​​​​​​​一​​​​​​​个​​​​​​​或​​​​​​​多​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​子​​​​​​​集​​​​​​​,用​​​​​​​来​​​​​​​在​​​​​​​节​​​​​​​点​​​​​​​发​​​​​​​生​​​​​​​故​​​​​​​障​​​​​​​时​​​​​​​运​​​​​​​行​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​出​​​​​​​现​​​​​​​在​​​​​​​「​​​​​​​Failover Domains」​​​​​​​的​​​​​​​下​​​​​​​级​​​​​​​列​​​​​​​表​​​​​​​里​​​​​​​。​​​​​​​使​​​​​​​用​​​​​​​右​​​​​​​栏​​​​​​​底​​​​​​​部​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​按​​​​​​​钮​​​​​​​(在​​​​​​​「​​​​​​​Properties」​​​​​​​下​​​​​​​面​​​​​​​),您​​​​​​​可​​​​​​​以​​​​​​​创​​​​​​​建​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​(当​​​​​​​选​​​​​​​定​​​​​​​了​​​​​​​「​​​​​​​Failover Domains」​​​​​​​项​​​​​​​时​​​​​​​)或​​​​​​​编​​​​​​​辑​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​的​​​​​​​属​​​​​​​性​​​​​​​(当​​​​​​​选​​​​​​​择​​​​​​​其​​​​​​​中​​​​​​​一​​​​​​​个​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​时​​​​​​​)。​​​​​​​
    • 「​​​​​​​Resources」​​​​​​​ — 配​​​​​​​置​​​​​​​用​​​​​​​于​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​共​​​​​​​享​​​​​​​资​​​​​​​源​​​​​​​。​​​​​​​共​​​​​​​享​​​​​​​资​​​​​​​源​​​​​​​由​​​​​​​可​​​​​​​用​​​​​​​于​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​的​​​​​​​任​​​​​​​何​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​、​​​​​​​IP 地​​​​​​​址​​​​​​​、​​​​​​​NFS 挂​​​​​​​载​​​​​​​和​​​​​​​导​​​​​​​出​​​​​​​以​​​​​​​及​​​​​​​用​​​​​​​户​​​​​​​创​​​​​​​建​​​​​​​的​​​​​​​脚​​​​​​​本​​​​​​​组​​​​​​​成​​​​​​​。​​​​​​​资​​​​​​​源​​​​​​​出​​​​​​​现​​​​​​​在​​​​​​​「​​​​​​​Resources」​​​​​​​的​​​​​​​下​​​​​​​级​​​​​​​列​​​​​​​表​​​​​​​里​​​​​​​。​​​​​​​使​​​​​​​用​​​​​​​右​​​​​​​栏​​​​​​​底​​​​​​​部​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​按​​​​​​​钮​​​​​​​(在​​​​​​​「​​​​​​​Properties」​​​​​​​下​​​​​​​面​​​​​​​),您​​​​​​​可​​​​​​​以​​​​​​​创​​​​​​​建​​​​​​​资​​​​​​​源​​​​​​​(当​​​​​​​选​​​​​​​住​​​​​​​了​​​​​​​「​​​​​​​Resources」​​​​​​​)或​​​​​​​编​​​​​​​辑​​​​​​​资​​​​​​​源​​​​​​​属​​​​​​​性​​​​​​​(但​​​​​​​选​​​​​​​中​​​​​​​了​​​​​​​其​​​​​​​中​​​​​​​一​​​​​​​个​​​​​​​资​​​​​​​源​​​​​​​)。​​​​​​​

      注意

      Cluster Configuration Tool 也​​​​​​​提​​​​​​​供​​​​​​​配​​​​​​​置​​​​​​​专​​​​​​​用​​​​​​​资​​​​​​​源​​​​​​​的​​​​​​​能​​​​​​​力​​​​​​​。​​​​​​​专​​​​​​​用​​​​​​​资​​​​​​​源​​​​​​​是​​​​​​​仅​​​​​​​用​​​​​​​于​​​​​​​一​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​资​​​​​​​源​​​​​​​。​​​​​​​您​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​图​​​​​​​形​​​​​​​用​​​​​​​户​​​​​​​界​​​​​​​面​​​​​​​里​​​​​​​的​​​​​​​ Service 组​​​​​​​件​​​​​​​里​​​​​​​配​​​​​​​置​​​​​​​专​​​​​​​用​​​​​​​资​​​​​​​源​​​​​​​。​​​​​​​
    • 「​​​​​​​Services」​​​​​​​ — 创​​​​​​​建​​​​​​​和​​​​​​​配​​​​​​​置​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​服​​​​​​​务​​​​​​​通​​​​​​​过​​​​​​​分​​​​​​​配​​​​​​​资​​​​​​​源​​​​​​​(共​​​​​​​享​​​​​​​的​​​​​​​或​​​​​​​专​​​​​​​用​​​​​​​的​​​​​​​)、​​​​​​​分​​​​​​​配​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​域​​​​​​​、​​​​​​​定​​​​​​​义​​​​​​​故​​​​​​​障​​​​​​​恢​​​​​​​复​​​​​​​策​​​​​​​略​​​​​​​来​​​​​​​进​​​​​​​行​​​​​​​配​​​​​​​置​​​​​​​。​​​​​​​服​​​​​​​务​​​​​​​出​​​​​​​现​​​​​​​在​​​​​​​「​​​​​​​Services」​​​​​​​的​​​​​​​下​​​​​​​级​​​​​​​列​​​​​​​表​​​​​​​里​​​​​​​。​​​​​​​使​​​​​​​用​​​​​​​右​​​​​​​栏​​​​​​​底​​​​​​​部​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​按​​​​​​​钮​​​​​​​(在​​​​​​​「​​​​​​​Properties」​​​​​​​下​​​​​​​面​​​​​​​),您​​​​​​​可​​​​​​​以​​​​​​​创​​​​​​​建​​​​​​​服​​​​​​​务​​​​​​​(当​​​​​​​选​​​​​​​住​​​​​​​「​​​​​​​Services」​​​​​​​)或​​​​​​​者​​​​​​​编​​​​​​​辑​​​​​​​服​​​​​​​务​​​​​​​属​​​​​​​性​​​​​​​(当​​​​​​​选​​​​​​​中​​​​​​​了​​​​​​​某​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​)。​​​​​​​

1.9.2.2. Cluster Status Tool

You can access the Cluster Status Tool (图 1.28 “Cluster Status Tool) through the Cluster Management tab in Cluster Administration GUI.
Cluster Status Tool

图 1.28. Cluster Status Tool

Cluster Status Tool 里​​​​​​​显​​​​​​​示​​​​​​​的​​​​​​​节​​​​​​​点​​​​​​​和​​​​​​​服​​​​​​​务​​​​​​​由​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​(/etc/cluster/cluster.conf)决​​​​​​​定​​​​​​​。​​​​​​​您​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​ Cluster Status Tool 来​​​​​​​启​​​​​​​用​​​​​​​、​​​​​​​禁​​​​​​​用​​​​​​​、​​​​​​​重​​​​​​​启​​​​​​​或​​​​​​​迁​​​​​​​移​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​

1.9.3. 命​​​​​​​令​​​​​​​行​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​

In addition to Conga and the system-config-cluster Cluster Administration GUI, command line tools are available for administering the cluster infrastructure and the high-availability service management components. The command line tools are used by the Cluster Administration GUI and init scripts supplied by Red Hat. 表 1.1 “命​​​​​​​令​​​​​​​行​​​​​​​工​​​​​​​具​​​​​​​” summarizes the command line tools.

表 1.1. 命​​​​​​​令​​​​​​​行​​​​​​​工​​​​​​​具​​​​​​​

命​​​​​​​令​​​​​​​行​​​​​​​工​​​​​​​具​​​​​​​ 用​​​​​​​于​​​​​​​ 描​​​​​​​述​​​​​​​
ccs_tool — 群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​系​​​​​​​统​​​​​​​工​​​​​​​具​​​​​​​ Cluster Infrastructure ccs_tool 是​​​​​​​一​​​​​​​个​​​​​​​制​​​​​​​作​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​的​​​​​​​在​​​​​​​线​​​​​​​更​​​​​​​新​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​。​​​​​​​它​​​​​​​提​​​​​​​供​​​​​​​创​​​​​​​建​​​​​​​和​​​​​​​修​​​​​​​改​​​​​​​群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​组​​​​​​​件​​​​​​​(如​​​​​​​:创​​​​​​​建​​​​​​​群​​​​​​​集​​​​​​​、​​​​​​​添​​​​​​​加​​​​​​​和​​​​​​​删​​​​​​​除​​​​​​​节​​​​​​​点​​​​​​​)的​​​​​​​能​​​​​​​力​​​​​​​。​​​​​​​关​​​​​​​于​​​​​​​这​​​​​​​个​​​​​​​工​​​​​​​具​​​​​​​的​​​​​​​更​​​​​​​多​​​​​​​信​​​​​​​息​​​​​​​,请​​​​​​​参​​​​​​​考​​​​​​​ ccs_tool(8) 帮​​​​​​​助​​​​​​​页​​​​​​​。​​​​​​​
cman_tool — 群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​ Cluster Infrastructure cman_tool 是​​​​​​​一​​​​​​​个​​​​​​​管​​​​​​​理​​​​​​​ CMAN 群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​。​​​​​​​它​​​​​​​提​​​​​​​供​​​​​​​加​​​​​​​入​​​​​​​群​​​​​​​集​​​​​​​、​​​​​​​离​​​​​​​开​​​​​​​群​​​​​​​集​​​​​​​、​​​​​​​杀​​​​​​​死​​​​​​​节​​​​​​​点​​​​​​​或​​​​​​​修​​​​​​​改​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​预​​​​​​​期​​​​​​​ quorum votes 的​​​​​​​功​​​​​​​能​​​​​​​。​​​​​​​关​​​​​​​于​​​​​​​这​​​​​​​个​​​​​​​工​​​​​​​具​​​​​​​的​​​​​​​更​​​​​​​多​​​​​​​信​​​​​​​息​​​​​​​,请​​​​​​​参​​​​​​​考​​​​​​​ cman_tool(8) 帮​​​​​​​助​​​​​​​页​​​​​​​。​​​​​​​
fence_tool — 保​​​​​​​护​​​​​​​工​​​​​​​具​​​​​​​ Cluster Infrastructure fence_tool 是​​​​​​​一​​​​​​​个​​​​​​​用​​​​​​​来​​​​​​​加​​​​​​​入​​​​​​​和​​​​​​​离​​​​​​​开​​​​​​​缺​​​​​​​省​​​​​​​保​​​​​​​护​​​​​​​域​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​。​​​​​​​具​​​​​​​体​​​​​​​来​​​​​​​说​​​​​​​,它​​​​​​​启​​​​​​​动​​​​​​​保​​​​​​​护​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​((fenced)来​​​​​​​加​​​​​​​入​​​​​​​域​​​​​​​以​​​​​​​及​​​​​​​终​​​​​​​止​​​​​​​ fenced 进​​​​​​​程​​​​​​​来​​​​​​​离​​​​​​​开​​​​​​​域​​​​​​​。​​​​​​​关​​​​​​​于​​​​​​​这​​​​​​​个​​​​​​​工​​​​​​​具​​​​​​​的​​​​​​​更​​​​​​​多​​​​​​​信​​​​​​​息​​​​​​​,请​​​​​​​参​​​​​​​考​​​​​​​ fence_tool(8) 帮​​​​​​​助​​​​​​​页​​​​​​​。​​​​​​​
clustat — 群​​​​​​​集​​​​​​​状​​​​​​​态​​​​​​​工​​​​​​​具​​​​​​​ 高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​组​​​​​​​件​​​​​​​ clustat 命​​​​​​​令​​​​​​​显​​​​​​​示​​​​​​​群​​​​​​​集​​​​​​​的​​​​​​​状​​​​​​​态​​​​​​​。​​​​​​​它​​​​​​​显​​​​​​​示​​​​​​​成​​​​​​​员​​​​​​​信​​​​​​​息​​​​​​​、​​​​​​​quorum 视​​​​​​​图​​​​​​​以​​​​​​​及​​​​​​​所​​​​​​​有​​​​​​​配​​​​​​​置​​​​​​​的​​​​​​​用​​​​​​​户​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​状​​​​​​​态​​​​​​​。​​​​​​​关​​​​​​​于​​​​​​​这​​​​​​​个​​​​​​​工​​​​​​​具​​​​​​​的​​​​​​​更​​​​​​​多​​​​​​​信​​​​​​​息​​​​​​​,请​​​​​​​参​​​​​​​考​​​​​​​ clustat(8) 帮​​​​​​​助​​​​​​​页​​​​​​​。​​​​​​​
clusvcadm — 群​​​​​​​集​​​​​​​用​​​​​​​户​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​ 高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​组​​​​​​​件​​​​​​​ clusvcadm 命​​​​​​​令​​​​​​​允​​​​​​​许​​​​​​​您​​​​​​​在​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​启​​​​​​​用​​​​​​​、​​​​​​​禁​​​​​​​用​​​​​​​、​​​​​​​迁​​​​​​​移​​​​​​​和​​​​​​​重​​​​​​​启​​​​​​​高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​关​​​​​​​于​​​​​​​这​​​​​​​个​​​​​​​工​​​​​​​具​​​​​​​的​​​​​​​更​​​​​​​多​​​​​​​信​​​​​​​息​​​​​​​,请​​​​​​​参​​​​​​​考​​​​​​​ clusvcadm(8) 帮​​​​​​​助​​​​​​​页​​​​​​​。​​​​​​​

1.10. Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​管​​​​​​​理​​​​​​​图​​​​​​​形​​​​​​​用​​​​​​​户​​​​​​​界​​​​​​​面​​​​​​​

本​​​​​​​节​​​​​​​提​​​​​​​供​​​​​​​对​​​​​​​ Red Hat Cluster Suite 里​​​​​​​可​​​​​​​用​​​​​​​的​​​​​​​ LVS 配​​​​​​​置​​​​​​​工​​​​​​​具​​​​​​​的​​​​​​​概​​​​​​​述​​​​​​​ — Piranha Configuration Tool。​​​​​​​Piranha Configuration Tool 是​​​​​​​一​​​​​​​个​​​​​​​ web 浏​​​​​​​览​​​​​​​器​​​​​​​图​​​​​​​形​​​​​​​用​​​​​​​户​​​​​​​界​​​​​​​面​​​​​​​(GUI),它​​​​​​​提​​​​​​​供​​​​​​​结​​​​​​​构​​​​​​​化​​​​​​​的​​​​​​​方​​​​​​​法​​​​​​​来​​​​​​​创​​​​​​​建​​​​​​​ LVS 配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​ — /etc/sysconfig/ha/lvs.cf。​​​​​​​
要​​​​​​​访​​​​​​​问​​​​​​​ Piranha Configuration Tool,您​​​​​​​需​​​​​​​要​​​​​​​在​​​​​​​活​​​​​​​动​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​上​​​​​​​运​​​​​​​行​​​​​​​ piranha-gui 服​​​​​​​务​​​​​​​。​​​​​​​您​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​本​​​​​​​地​​​​​​​访​​​​​​​问​​​​​​​ Piranha Configuration Tool 或​​​​​​​通​​​​​​​过​​​​​​​ web 浏​​​​​​​览​​​​​​​器​​​​​​​进​​​​​​​行​​​​​​​远​​​​​​​程​​​​​​​访​​​​​​​问​​​​​​​。​​​​​​​本​​​​​​​地​​​​​​​访​​​​​​​问​​​​​​​您​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​链​​​​​​​接​​​​​​​: http://localhost:3636。​​​​​​​如​​​​​​​果​​​​​​​要​​​​​​​从​​​​​​​远​​​​​​​程​​​​​​​访​​​​​​​问​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​用​​​​​​​主​​​​​​​机​​​​​​​名​​​​​​​或​​​​​​​者​​​​​​​真​​​​​​​实​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​后​​​​​​​面​​​​​​​接​​​​​​​着​​​​​​​ :3636。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​从​​​​​​​远​​​​​​​程​​​​​​​访​​​​​​​问​​​​​​​ Piranha Configuration Tool,您​​​​​​​需​​​​​​​要​​​​​​​以​​​​​​​根​​​​​​​用​​​​​​​户​​​​​​​身​​​​​​​份​​​​​​​用​​​​​​​ ssh 连​​​​​​​接​​​​​​​活​​​​​​​动​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​。​​​​​​​
Starting the Piranha Configuration Tool causes the Piranha Configuration Tool welcome page to be displayed (refer to 图 1.29 “The Welcome Panel”). Logging in to the welcome page provides access to the four main screens or panels: CONTROL/MONITORING, GLOBAL SETTINGS, REDUNDANCY, and VIRTUAL SERVERS. In addition, the VIRTUAL SERVERS panel contains four subsections. The CONTROL/MONITORING panel is the first panel displayed after you log in at the welcome screen.
The Welcome Panel

图 1.29. The Welcome Panel

下​​​​​​​面​​​​​​​的​​​​​​​内​​​​​​​容​​​​​​​提​​​​​​​供​​​​​​​了​​​​​​​对​​​​​​​ Piranha Configuration Tool 配​​​​​​​置​​​​​​​页​​​​​​​面​​​​​​​的​​​​​​​简​​​​​​​短​​​​​​​描​​​​​​​述​​​​​​​。​​​​​​​

1.10.1. CONTROL/MONITORING

「​​​​​​​CONTROL/MONITORING」​​​​​​​面​​​​​​​板​​​​​​​显​​​​​​​示​​​​​​​运​​​​​​​行​​​​​​​状​​​​​​​态​​​​​​​。​​​​​​​它​​​​​​​显​​​​​​​示​​​​​​​ pulse 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​、​​​​​​​LVS 路​​​​​​​由​​​​​​​表​​​​​​​以​​​​​​​及​​​​​​​ LVS 所​​​​​​​产​​​​​​​生​​​​​​​的​​​​​​​ nanny 进​​​​​​​程​​​​​​​的​​​​​​​状​​​​​​​态​​​​​​​。​​​​​​​
The CONTROL/MONITORING Panel

图 1.30. The CONTROL/MONITORING Panel

Auto update
让​​​​​​​状​​​​​​​态​​​​​​​显​​​​​​​示​​​​​​​按​​​​​​​照​​​​​​​「​​​​​​​Update frequency in seconds」​​​​​​​文​​​​​​​本​​​​​​​框​​​​​​​里​​​​​​​设​​​​​​​置​​​​​​​时​​​​​​​间​​​​​​​间​​​​​​​隔​​​​​​​(缺​​​​​​​省​​​​​​​为​​​​​​​ 10 秒​​​​​​​)自​​​​​​​动​​​​​​​进​​​​​​​行​​​​​​​更​​​​​​​新​​​​​​​。​​​​​​​
我​​​​​​​们​​​​​​​不​​​​​​​推​​​​​​​荐​​​​​​​将​​​​​​​自​​​​​​​动​​​​​​​更​​​​​​​新​​​​​​​的​​​​​​​时​​​​​​​间​​​​​​​间​​​​​​​隔​​​​​​​设​​​​​​​置​​​​​​​为​​​​​​​少​​​​​​​于​​​​​​​ 10 秒​​​​​​​。​​​​​​​这​​​​​​​样​​​​​​​使​​​​​​​重​​​​​​​新​​​​​​​配​​​​​​​置​​​​​​​「​​​​​​​Auto update」​​​​​​​时​​​​​​​间​​​​​​​间​​​​​​​隔​​​​​​​变​​​​​​​得​​​​​​​困​​​​​​​难​​​​​​​,因​​​​​​​为​​​​​​​页​​​​​​​面​​​​​​​会​​​​​​​过​​​​​​​于​​​​​​​频​​​​​​​繁​​​​​​​地​​​​​​​更​​​​​​​新​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​遇​​​​​​​到​​​​​​​这​​​​​​​个​​​​​​​问​​​​​​​题​​​​​​​,只​​​​​​​要​​​​​​​简​​​​​​​单​​​​​​​地​​​​​​​点​​​​​​​击​​​​​​​另​​​​​​​外​​​​​​​一​​​​​​​个​​​​​​​面​​​​​​​板​​​​​​​并​​​​​​​回​​​​​​​到​​​​​​​「​​​​​​​CONTROL/MONITORING」​​​​​​​就​​​​​​​可​​​​​​​以​​​​​​​了​​​​​​​。​​​​​​​
Update information now
提​​​​​​​供​​​​​​​对​​​​​​​状​​​​​​​态​​​​​​​信​​​​​​​息​​​​​​​的​​​​​​​手​​​​​​​工​​​​​​​更​​​​​​​新​​​​​​​。​​​​​​​
CHANGE PASSWORD
点​​​​​​​击​​​​​​​这​​​​​​​个​​​​​​​按​​​​​​​钮​​​​​​​会​​​​​​​使​​​​​​​您​​​​​​​进​​​​​​​入​​​​​​​一​​​​​​​个​​​​​​​帮​​​​​​​助​​​​​​​屏​​​​​​​幕​​​​​​​,上​​​​​​​面​​​​​​​有​​​​​​​如​​​​​​​何​​​​​​​修​​​​​​​改​​​​​​​ Piranha Configuration Tool 的​​​​​​​管​​​​​​​理​​​​​​​密​​​​​​​码​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​

1.10.2. GLOBAL SETTINGS

The GLOBAL SETTINGS panel is where the LVS administrator defines the networking details for the primary LVS router's public and private network interfaces.
The GLOBAL SETTINGS Panel

图 1.31. The GLOBAL SETTINGS Panel

The top half of this panel sets up the primary LVS router's public and private network interfaces.
Primary server public IP
主​​​​​​​ LVS 节​​​​​​​点​​​​​​​的​​​​​​​公​​​​​​​用​​​​​​​真​​​​​​​实​​​​​​​ IP 地​​​​​​​址​​​​​​​。​​​​​​​
Primary server private IP
主​​​​​​​ LVS 节​​​​​​​点​​​​​​​的​​​​​​​替​​​​​​​代​​​​​​​网​​​​​​​络​​​​​​​接​​​​​​​口​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​ IP 地​​​​​​​址​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​地​​​​​​​址​​​​​​​仅​​​​​​​用​​​​​​​于​​​​​​​与​​​​​​​备​​​​​​​份​​​​​​​路​​​​​​​由​​​​​​​器​​​​​​​之​​​​​​​间​​​​​​​的​​​​​​​ heartbeat 通​​​​​​​道​​​​​​​。​​​​​​​
Use network type
选​​​​​​​择​​​​​​​ NAT 路​​​​​​​由​​​​​​​
The next three fields are specifically for the NAT router's virtual network interface connected the private network with the real servers.
NAT Router IP
这​​​​​​​个​​​​​​​文​​​​​​​本​​​​​​​字​​​​​​​段​​​​​​​里​​​​​​​是​​​​​​​专​​​​​​​用​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​应​​​​​​​该​​​​​​​作​​​​​​​为​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​网​​​​​​​关​​​​​​​。​​​​​​​
NAT Router netmask
If the NAT router's floating IP needs a particular netmask, select it from drop-down list.
NAT Router device
定​​​​​​​义​​​​​​​用​​​​​​​于​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​的​​​​​​​网​​​​​​​络​​​​​​​接​​​​​​​口​​​​​​​的​​​​​​​设​​​​​​​备​​​​​​​名​​​​​​​称​​​​​​​,如​​​​​​​ eth1:1。​​​​​​​

1.10.3. REDUNDANCY

「​​​​​​​REDUNDANCY」​​​​​​​面​​​​​​​板​​​​​​​允​​​​​​​许​​​​​​​您​​​​​​​配​​​​​​​置​​​​​​​备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​节​​​​​​​点​​​​​​​并​​​​​​​设​​​​​​​置​​​​​​​不​​​​​​​同​​​​​​​的​​​​​​​ heartbeat 监​​​​​​​控​​​​​​​选​​​​​​​项​​​​​​​。​​​​​​​
The REDUNDANCY Panel

图 1.32. The REDUNDANCY Panel

Redundant server public IP
备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​的​​​​​​​公​​​​​​​用​​​​​​​的​​​​​​​真​​​​​​​实​​​​​​​ IP 地​​​​​​​址​​​​​​​。​​​​​​​
Redundant server private IP
The backup router's private real IP address.
面​​​​​​​板​​​​​​​的​​​​​​​其​​​​​​​他​​​​​​​部​​​​​​​分​​​​​​​包​​​​​​​括​​​​​​​配​​​​​​​置​​​​​​​ heartbeat 通​​​​​​​道​​​​​​​,备​​​​​​​份​​​​​​​节​​​​​​​点​​​​​​​用​​​​​​​它​​​​​​​来​​​​​​​监​​​​​​​控​​​​​​​主​​​​​​​节​​​​​​​点​​​​​​​的​​​​​​​故​​​​​​​障​​​​​​​情​​​​​​​况​​​​​​​。​​​​​​​
Heartbeat Interval (seconds)
设​​​​​​​置​​​​​​​ heartbeat 之​​​​​​​间​​​​​​​间​​​​​​​隔​​​​​​​的​​​​​​​秒​​​​​​​数​​​​​​​ — 备​​​​​​​份​​​​​​​节​​​​​​​点​​​​​​​检​​​​​​​查​​​​​​​主​​​​​​​ LVS 节​​​​​​​点​​​​​​​功​​​​​​​能​​​​​​​性​​​​​​​状​​​​​​​态​​​​​​​的​​​​​​​时​​​​​​​间​​​​​​​间​​​​​​​隔​​​​​​​。​​​​​​​
Assume dead after (seconds)
如​​​​​​​果​​​​​​​主​​​​​​​ LVS 节​​​​​​​点​​​​​​​在​​​​​​​这​​​​​​​段​​​​​​​时​​​​​​​间​​​​​​​内​​​​​​​没​​​​​​​有​​​​​​​响​​​​​​​应​​​​​​​,备​​​​​​​份​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​节​​​​​​​点​​​​​​​将​​​​​​​初​​​​​​​始​​​​​​​化​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​。​​​​​​​
Heartbeat runs on port
设​​​​​​​置​​​​​​​ heartbeat 与​​​​​​​主​​​​​​​ LVS 节​​​​​​​点​​​​​​​通​​​​​​​信​​​​​​​的​​​​​​​端​​​​​​​口​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​不​​​​​​​填​​​​​​​写​​​​​​​这​​​​​​​个​​​​​​​字​​​​​​​段​​​​​​​,其​​​​​​​省​​​​​​​省​​​​​​​值​​​​​​​是​​​​​​​ 539。​​​​​​​

1.10.4. VIRTUAL SERVERS

「​​​​​​​VIRTUAL SERVERS」​​​​​​​面​​​​​​​板​​​​​​​显​​​​​​​示​​​​​​​每​​​​​​​个​​​​​​​当​​​​​​​前​​​​​​​定​​​​​​​义​​​​​​​的​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​表​​​​​​​里​​​​​​​的​​​​​​​每​​​​​​​个​​​​​​​条​​​​​​​目​​​​​​​都​​​​​​​显​​​​​​​示​​​​​​​了​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​状​​​​​​​态​​​​​​​、​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​名​​​​​​​、​​​​​​​分​​​​​​​配​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​、​​​​​​​虚​​​​​​​拟​​​​​​​ IP 的​​​​​​​掩​​​​​​​码​​​​​​​、​​​​​​​服​​​​​​​务​​​​​​​端​​​​​​​口​​​​​​​、​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​协​​​​​​​议​​​​​​​以​​​​​​​及​​​​​​​虚​​​​​​​拟​​​​​​​设​​​​​​​备​​​​​​​接​​​​​​​口​​​​​​​。​​​​​​​
The VIRTUAL SERVERS Panel

图 1.33. The VIRTUAL SERVERS Panel

「​​​​​​​VIRTUAL SERVERS」​​​​​​​面​​​​​​​将​​​​​​​里​​​​​​​显​​​​​​​示​​​​​​​的​​​​​​​每​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​都​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​接​​​​​​​下​​​​​​​来​​​​​​​的​​​​​​​屏​​​​​​​幕​​​​​​​或​​​​​​​分​​​​​​​支​​​​​​​(subsections)里​​​​​​​进​​​​​​​行​​​​​​​配​​​​​​​置​​​​​​​。​​​​​​​
点​​​​​​​击​​​​​​​「​​​​​​​ADD」​​​​​​​按​​​​​​​钮​​​​​​​可​​​​​​​以​​​​​​​添​​​​​​​加​​​​​​​一​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​要​​​​​​​删​​​​​​​除​​​​​​​某​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​,可​​​​​​​以​​​​​​​选​​​​​​​中​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​旁​​​​​​​的​​​​​​​单​​​​​​​选​​​​​​​按​​​​​​​钮​​​​​​​并​​​​​​​点​​​​​​​击​​​​​​​「​​​​​​​DELETE」​​​​​​​按​​​​​​​钮​​​​​​​。​​​​​​​
要​​​​​​​启​​​​​​​用​​​​​​​或​​​​​​​禁​​​​​​​用​​​​​​​表​​​​​​​里​​​​​​​的​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​,选​​​​​​​中​​​​​​​单​​​​​​​选​​​​​​​按​​​​​​​钮​​​​​​​并​​​​​​​点​​​​​​​击​​​​​​​「​​​​​​​(DE)ACTIVATE」​​​​​​​按​​​​​​​钮​​​​​​​。​​​​​​​
添​​​​​​​加​​​​​​​了​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​后​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​选​​​​​​​中​​​​​​​其​​​​​​​左​​​​​​​边​​​​​​​的​​​​​​​单​​​​​​​选​​​​​​​按​​​​​​​钮​​​​​​​并​​​​​​​点​​​​​​​击​​​​​​​「​​​​​​​EDITEDIT」​​​​​​​按​​​​​​​钮​​​​​​​进​​​​​​​入​​​​​​​「​​​​​​​VIRTUAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​来​​​​​​​进​​​​​​​行​​​​​​​配​​​​​​​置​​​​​​​。​​​​​​​

1.10.4.1. 「​​​​​​​VIRTUAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​

The VIRTUAL SERVER subsection panel shown in 图 1.34 “The VIRTUAL SERVERS Subsection” allows you to configure an individual virtual server. Links to subsections related specifically to this virtual server are located along the top of the page. But before configuring any of the subsections related to this virtual server, complete this page and click on the ACCEPT button.
The VIRTUAL SERVERS Subsection

图 1.34. The VIRTUAL SERVERS Subsection

Name
标​​​​​​​识​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​描​​​​​​​述​​​​​​​性​​​​​​​名​​​​​​​称​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​名​​​​​​​称​​​​​​​不​​​​​​​是​​​​​​​机​​​​​​​器​​​​​​​的​​​​​​​主​​​​​​​机​​​​​​​名​​​​​​​,所​​​​​​​以​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​描​​​​​​​述​​​​​​​性​​​​​​​且​​​​​​​易​​​​​​​于​​​​​​​辨​​​​​​​别​​​​​​​的​​​​​​​文​​​​​​​字​​​​​​​。​​​​​​​您​​​​​​​甚​​​​​​​至​​​​​​​可​​​​​​​以​​​​​​​引​​​​​​​用​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​协​​​​​​​议​​​​​​​,如​​​​​​​ HTTP。​​​​​​​
Application port
应​​​​​​​用​​​​​​​程​​​​​​​序​​​​​​​将​​​​​​​侦​​​​​​​听​​​​​​​的​​​​​​​端​​​​​​​口​​​​​​​。​​​​​​​
Protocol
在​​​​​​​下​​​​​​​拉​​​​​​​菜​​​​​​​单​​​​​​​里​​​​​​​提​​​​​​​供​​​​​​​ UDP 或​​​​​​​ TCP 选​​​​​​​项​​​​​​​。​​​​​​​
Virtual IP Address
The virtual server's floating IP address.
Virtual IP Network Mask
可​​​​​​​在​​​​​​​下​​​​​​​拉​​​​​​​菜​​​​​​​单​​​​​​​里​​​​​​​选​​​​​​​择​​​​​​​的​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​掩​​​​​​​码​​​​​​​。​​​​​​​
Firewall Mark
捆​​​​​​​绑​​​​​​​多​​​​​​​端​​​​​​​口​​​​​​​的​​​​​​​协​​​​​​​议​​​​​​​或​​​​​​​创​​​​​​​建​​​​​​​使​​​​​​​用​​​​​​​独​​​​​​​立​​​​​​​但​​​​​​​相​​​​​​​关​​​​​​​的​​​​​​​协​​​​​​​议​​​​​​​的​​​​​​​多​​​​​​​端​​​​​​​口​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​时​​​​​​​,输​​​​​​​入​​​​​​​的​​​​​​​防​​​​​​​火​​​​​​​墙​​​​​​​掩​​​​​​​码​​​​​​​整​​​​​​​型​​​​​​​数​​​​​​​值​​​​​​​。​​​​​​​
Device
「​​​​​​​Virtual IP Address」​​​​​​​里​​​​​​​定​​​​​​​义​​​​​​​的​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​绑​​​​​​​定​​​​​​​的​​​​​​​网​​​​​​​络​​​​​​​设​​​​​​​备​​​​​​​名​​​​​​​称​​​​​​​。​​​​​​​
您​​​​​​​应​​​​​​​该​​​​​​​为​​​​​​​连​​​​​​​接​​​​​​​至​​​​​​​公​​​​​​​用​​​​​​​网​​​​​​​络​​​​​​​的​​​​​​​以​​​​​​​太​​​​​​​接​​​​​​​口​​​​​​​的​​​​​​​公​​​​​​​用​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​起​​​​​​​别​​​​​​​名​​​​​​​。​​​​​​​
Re-entry Time
服​​​​​​​务​​​​​​​器​​​​​​​发​​​​​​​生​​​​​​​故​​​​​​​障​​​​​​​后​​​​​​​,活​​​​​​​动​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​试​​​​​​​图​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​前​​​​​​​等​​​​​​​待​​​​​​​的​​​​​​​时​​​​​​​间​​​​​​​(以​​​​​​​秒​​​​​​​为​​​​​​​单​​​​​​​位​​​​​​​,整​​​​​​​型​​​​​​​数​​​​​​​值​​​​​​​)。​​​​​​​
Service Timeout
定​​​​​​​义​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​被​​​​​​​认​​​​​​​为​​​​​​​发​​​​​​​生​​​​​​​了​​​​​​​故​​​​​​​障​​​​​​​或​​​​​​​不​​​​​​​可​​​​​​​用​​​​​​​之​​​​​​​前​​​​​​​的​​​​​​​秒​​​​​​​数​​​​​​​的​​​​​​​整​​​​​​​型​​​​​​​数​​​​​​​值​​​​​​​。​​​​​​​
Quiesce server
当​​​​​​​「​​​​​​​Quiesce server」​​​​​​​单​​​​​​​选​​​​​​​按​​​​​​​钮​​​​​​​被​​​​​​​选​​​​​​​中​​​​​​​时​​​​​​​,新​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​节​​​​​​​点​​​​​​​上​​​​​​​线​​​​​​​时​​​​​​​,最​​​​​​​少​​​​​​​连​​​​​​​接​​​​​​​表​​​​​​​被​​​​​​​重​​​​​​​置​​​​​​​为​​​​​​​ 0,活​​​​​​​动​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​将​​​​​​​发​​​​​​​送​​​​​​​请​​​​​​​求​​​​​​​,就​​​​​​​好​​​​​​​像​​​​​​​所​​​​​​​有​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​都​​​​​​​是​​​​​​​刚​​​​​​​加​​​​​​​入​​​​​​​群​​​​​​​集​​​​​​​一​​​​​​​样​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​选​​​​​​​项​​​​​​​可​​​​​​​以​​​​​​​避​​​​​​​免​​​​​​​当​​​​​​​大​​​​​​​量​​​​​​​的​​​​​​​连​​​​​​​接​​​​​​​进​​​​​​​入​​​​​​​群​​​​​​​集​​​​​​​时​​​​​​​,新​​​​​​​加​​​​​​​入​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​超​​​​​​​载​​​​​​​。​​​​​​​
Load monitoring tool
LVS 路​​​​​​​由​​​​​​​器​​​​​​​可​​​​​​​以​​​​​​​用​​​​​​​ rup 或​​​​​​​ ruptime 监​​​​​​​控​​​​​​​不​​​​​​​同​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​的​​​​​​​负​​​​​​​载​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​从​​​​​​​下​​​​​​​拉​​​​​​​菜​​​​​​​单​​​​​​​里​​​​​​​选​​​​​​​择​​​​​​​了​​​​​​​ rup,那​​​​​​​么​​​​​​​每​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​都​​​​​​​必​​​​​​​须​​​​​​​运​​​​​​​行​​​​​​​ rstatd 服​​​​​​​务​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​选​​​​​​​择​​​​​​​了​​​​​​​ ruptime,每​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​都​​​​​​​必​​​​​​​须​​​​​​​运​​​​​​​行​​​​​​​ rwhod 服​​​​​​​务​​​​​​​。​​​​​​​
Scheduling
下​​​​​​​拉​​​​​​​菜​​​​​​​单​​​​​​​里​​​​​​​的​​​​​​​首​​​​​​​选​​​​​​​调​​​​​​​度​​​​​​​算​​​​​​​法​​​​​​​。​​​​​​​缺​​​​​​​省​​​​​​​是​​​​​​​权​​​​​​​重​​​​​​​最​​​​​​​少​​​​​​​连​​​​​​​接​​​​​​​(Weighted least-connection)算​​​​​​​法​​​​​​​。​​​​​​​
Persistence
在​​​​​​​处​​​​​​​理​​​​​​​客​​​​​​​户​​​​​​​事​​​​​​​务​​​​​​​时​​​​​​​,如​​​​​​​果​​​​​​​您​​​​​​​需​​​​​​​要​​​​​​​与​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​持​​​​​​​久​​​​​​​的​​​​​​​连​​​​​​​接​​​​​​​,可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​属​​​​​​​性​​​​​​​。​​​​​​​在​​​​​​​这​​​​​​​个​​​​​​​文​​​​​​​本​​​​​​​字​​​​​​​段​​​​​​​里​​​​​​​指​​​​​​​定​​​​​​​连​​​​​​​接​​​​​​​超​​​​​​​时​​​​​​​前​​​​​​​允​​​​​​​许​​​​​​​的​​​​​​​不​​​​​​​活​​​​​​​动​​​​​​​时​​​​​​​间​​​​​​​(秒​​​​​​​)。​​​​​​​
Persistence Network Mask
要​​​​​​​将​​​​​​​持​​​​​​​久​​​​​​​性​​​​​​​限​​​​​​​制​​​​​​​到​​​​​​​特​​​​​​​定​​​​​​​的​​​​​​​子​​​​​​​网​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​从​​​​​​​下​​​​​​​拉​​​​​​​菜​​​​​​​单​​​​​​​里​​​​​​​选​​​​​​​择​​​​​​​合​​​​​​​适​​​​​​​的​​​​​​​网​​​​​​​络​​​​​​​掩​​​​​​​码​​​​​​​。​​​​​​​

1.10.4.2. 「​​​​​​​REAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​

点​​​​​​​击​​​​​​​面​​​​​​​板​​​​​​​顶​​​​​​​部​​​​​​​的​​​​​​​「​​​​​​​REAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​链​​​​​​​接​​​​​​​将​​​​​​​显​​​​​​​示​​​​​​​「​​​​​​​EDIT REAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​。​​​​​​​
The REAL SERVER Subsection

图 1.35. The REAL SERVER Subsection

Click the ADD button to add a new server. To delete an existing server, select the radio button beside it and click the DELETE button. Click the EDIT button to load the EDIT REAL SERVER panel, as seen in 图 1.36 “The REAL SERVER Configuration Panel”.
The REAL SERVER Configuration Panel

图 1.36. The REAL SERVER Configuration Panel

这​​​​​​​个​​​​​​​面​​​​​​​板​​​​​​​由​​​​​​​ 3 个​​​​​​​字​​​​​​​段​​​​​​​组​​​​​​​成​​​​​​​:
Name
真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​描​​​​​​​述​​​​​​​性​​​​​​​名​​​​​​​称​​​​​​​。​​​​​​​

注意

这​​​​​​​个​​​​​​​名​​​​​​​称​​​​​​​不​​​​​​​是​​​​​​​机​​​​​​​器​​​​​​​的​​​​​​​主​​​​​​​机​​​​​​​名​​​​​​​,它​​​​​​​应​​​​​​​该​​​​​​​具​​​​​​​有​​​​​​​描​​​​​​​述​​​​​​​性​​​​​​​且​​​​​​​易​​​​​​​于​​​​​​​辨​​​​​​​认​​​​​​​。​​​​​​​
Address
The real server's IP address. Since the listening port is already specified for the associated virtual server, do not add a port number.
Weight
An integer value indicating this host's capacity relative to that of other hosts in the pool. The value can be arbitrary, but treat it as a ratio in relation to other real servers.

1.10.4.3. EDIT MONITORING SCRIPTS Subsection

点​​​​​​​击​​​​​​​页​​​​​​​面​​​​​​​顶​​​​​​​部​​​​​​​的​​​​​​​「​​​​​​​MONITORING SCRIPTS」​​​​​​​链​​​​​​​接​​​​​​​。​​​​​​​「​​​​​​​EDIT MONITORING SCRIPTS」​​​​​​​分​​​​​​​支​​​​​​​允​​​​​​​许​​​​​​​管​​​​​​​理​​​​​​​员​​​​​​​指​​​​​​​定​​​​​​​一​​​​​​​个​​​​​​​ send/expect 字​​​​​​​符​​​​​​​串​​​​​​​序​​​​​​​列​​​​​​​来​​​​​​​验​​​​​​​证​​​​​​​虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​服​​​​​​​务​​​​​​​在​​​​​​​每​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​正​​​​​​​常​​​​​​​运​​​​​​​行​​​​​​​。​​​​​​​管​​​​​​​理​​​​​​​员​​​​​​​也​​​​​​​可​​​​​​​以​​​​​​​在​​​​​​​这​​​​​​​里​​​​​​​定​​​​​​​义​​​​​​​检​​​​​​​查​​​​​​​需​​​​​​​要​​​​​​​动​​​​​​​态​​​​​​​数​​​​​​​据​​​​​​​更​​​​​​​新​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​自​​​​​​​定​​​​​​​义​​​​​​​脚​​​​​​​本​​​​​​​。​​​​​​​
The EDIT MONITORING SCRIPTS Subsection

图 1.37. The EDIT MONITORING SCRIPTS Subsection

Sending Program
对​​​​​​​于​​​​​​​更​​​​​​​高​​​​​​​级​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​验​​​​​​​证​​​​​​​,您​​​​​​​可​​​​​​​以​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​字​​​​​​​段​​​​​​​来​​​​​​​指​​​​​​​定​​​​​​​服​​​​​​​务​​​​​​​检​​​​​​​查​​​​​​​脚​​​​​​​本​​​​​​​的​​​​​​​路​​​​​​​径​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​功​​​​​​​能​​​​​​​对​​​​​​​于​​​​​​​需​​​​​​​要​​​​​​​动​​​​​​​态​​​​​​​更​​​​​​​新​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​来​​​​​​​说​​​​​​​尤​​​​​​​其​​​​​​​有​​​​​​​用​​​​​​​,如​​​​​​​ HTTPS 或​​​​​​​ SSL。​​​​​​​
要​​​​​​​使​​​​​​​用​​​​​​​这​​​​​​​个​​​​​​​功​​​​​​​能​​​​​​​,您​​​​​​​必​​​​​​​须​​​​​​​编​​​​​​​写​​​​​​​一​​​​​​​个​​​​​​​返​​​​​​​回​​​​​​​文​​​​​​​本​​​​​​​型​​​​​​​响​​​​​​​应​​​​​​​的​​​​​​​脚​​​​​​​本​​​​​​​,将​​​​​​​它​​​​​​​设​​​​​​​置​​​​​​​可​​​​​​​执​​​​​​​行​​​​​​​文​​​​​​​件​​​​​​​,并​​​​​​​在​​​​​​​「​​​​​​​Sending Program」​​​​​​​里​​​​​​​输​​​​​​​入​​​​​​​其​​​​​​​路​​​​​​​径​​​​​​​。​​​​​​​

注意

如​​​​​​​果​​​​​​​在​​​​​​​「​​​​​​​Sending Program」​​​​​​​字​​​​​​​段​​​​​​​里​​​​​​​输​​​​​​​入​​​​​​​了​​​​​​​外​​​​​​​部​​​​​​​程​​​​​​​序​​​​​​​,「​​​​​​​Send」​​​​​​​将​​​​​​​被​​​​​​​忽​​​​​​​略​​​​​​​。​​​​​​​
Send
nanny 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​将​​​​​​​这​​​​​​​个​​​​​​​字​​​​​​​段​​​​​​​里​​​​​​​的​​​​​​​字​​​​​​​符​​​​​​​串​​​​​​​发​​​​​​​送​​​​​​​到​​​​​​​每​​​​​​​个​​​​​​​真​​​​​​​实​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​字​​​​​​​段​​​​​​​里​​​​​​​缺​​​​​​​省​​​​​​​是​​​​​​​ HTTP 信​​​​​​​息​​​​​​​。​​​​​​​您​​​​​​​可​​​​​​​以​​​​​​​根​​​​​​​据​​​​​​​需​​​​​​​要​​​​​​​修​​​​​​​改​​​​​​​它​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​让​​​​​​​这​​​​​​​个​​​​​​​字​​​​​​​段​​​​​​​为​​​​​​​空​​​​​​​,nanny 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​将​​​​​​​试​​​​​​​图​​​​​​​打​​​​​​​开​​​​​​​服​​​​​​​务​​​​​​​端​​​​​​​口​​​​​​​,如​​​​​​​果​​​​​​​成​​​​​​​功​​​​​​​则​​​​​​​假​​​​​​​设​​​​​​​服​​​​​​​务​​​​​​​正​​​​​​​在​​​​​​​运​​​​​​​行​​​​​​​。​​​​​​​
这​​​​​​​个​​​​​​​字​​​​​​​段​​​​​​​里​​​​​​​只​​​​​​​允​​​​​​​许​​​​​​​输​​​​​​​入​​​​​​​一​​​​​​​个​​​​​​​发​​​​​​​送​​​​​​​序​​​​​​​列​​​​​​​,且​​​​​​​它​​​​​​​只​​​​​​​能​​​​​​​包​​​​​​​含​​​​​​​可​​​​​​​打​​​​​​​印​​​​​​​的​​​​​​​、​​​​​​​ASCII 字​​​​​​​符​​​​​​​串​​​​​​​以​​​​​​​及​​​​​​​下​​​​​​​面​​​​​​​的​​​​​​​转​​​​​​​义​​​​​​​符​​​​​​​:
  • \n 新​​​​​​​行​​​​​​​。​​​​​​​
  • \r 回​​​​​​​车​​​​​​​。​​​​​​​
  • \t 制​​​​​​​表​​​​​​​符​​​​​​​。​​​​​​​
  • \ 将​​​​​​​下​​​​​​​一​​​​​​​个​​​​​​​字​​​​​​​符​​​​​​​转​​​​​​​义​​​​​​​。​​​​​​​
Expect
如​​​​​​​果​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​正​​​​​​​常​​​​​​​运​​​​​​​行​​​​​​​,它​​​​​​​应​​​​​​​该​​​​​​​返​​​​​​​回​​​​​​​的​​​​​​​文​​​​​​​本​​​​​​​型​​​​​​​响​​​​​​​应​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​您​​​​​​​编​​​​​​​写​​​​​​​了​​​​​​​自​​​​​​​己​​​​​​​的​​​​​​​发​​​​​​​送​​​​​​​程​​​​​​​序​​​​​​​,在​​​​​​​此​​​​​​​输​​​​​​​入​​​​​​​您​​​​​​​设​​​​​​​定​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​返​​​​​​​回​​​​​​​的​​​​​​​响​​​​​​​应​​​​​​​。​​​​​​​


[1] 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​是​​​​​​​一​​​​​​​个​​​​​​​侦​​​​​​​听​​​​​​​特​​​​​​​定​​​​​​​的​​​​​​​虚​​​​​​​拟​​​​​​​ IP 地​​​​​​​址​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​

第 2 章 Red Hat Cluster Suite 组​​​​​​​件​​​​​​​摘​​​​​​​要​​​​​​​

本​​​​​​​章​​​​​​​提​​​​​​​供​​​​​​​了​​​​​​​ Red Hat Cluster Suite 组​​​​​​​件​​​​​​​的​​​​​​​摘​​​​​​​要​​​​​​​,它​​​​​​​由​​​​​​​下​​​​​​​面​​​​​​​的​​​​​​​章​​​​​​​节​​​​​​​组​​​​​​​成​​​​​​​:

2.1. 群​​​​​​​集​​​​​​​组​​​​​​​件​​​​​​​

表 2.1. Red Hat Cluster Suite 软​​​​​​​件​​​​​​​子​​​​​​​系​​​​​​​统​​​​​​​组​​​​​​​件​​​​​​​

功​​​​​​​能​​​​​​​ 组​​​​​​​件​​​​​​​ 描​​​​​​​述​​​​​​​
Conga luci 远​​​​​​​程​​​​​​​管​​​​​​​理​​​​​​​系​​​​​​​统​​​​​​​ - 管​​​​​​​理​​​​​​​工​​​​​​​作​​​​​​​站​​​​​​​。​​​​​​​
ricci 远​​​​​​​程​​​​​​​管​​​​​​​理​​​​​​​系​​​​​​​统​​​​​​​ - 被​​​​​​​管​​​​​​​理​​​​​​​的​​​​​​​工​​​​​​​作​​​​​​​站​​​​​​​。​​​​​​​
Cluster Configuration Tool system-config-cluster 以​​​​​​​图​​​​​​​形​​​​​​​方​​​​​​​式​​​​​​​管​​​​​​​理​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​的​​​​​​​命​​​​​​​令​​​​​​​
群​​​​​​​集​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​(Cluster Logical Volume Manager,CLVM) clvmd 在​​​​​​​群​​​​​​​集​​​​​​​中​​​​​​​发​​​​​​​布​​​​​​​ LVM 元​​​​​​​数​​​​​​​据​​​​​​​更​​​​​​​新​​​​​​​的​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​。​​​​​​​它​​​​​​​必​​​​​​​须​​​​​​​在​​​​​​​群​​​​​​​集​​​​​​​中​​​​​​​的​​​​​​​所​​​​​​​有​​​​​​​节​​​​​​​点​​​​​​​中​​​​​​​运​​​​​​​行​​​​​​​,如​​​​​​​果​​​​​​​群​​​​​​​集​​​​​​​中​​​​​​​的​​​​​​​某​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​没​​​​​​​有​​​​​​​运​​​​​​​行​​​​​​​此​​​​​​​时​​​​​​​间​​​​​​​进​​​​​​​程​​​​​​​则​​​​​​​会​​​​​​​提​​​​​​​示​​​​​​​错​​​​​​​误​​​​​​​。​​​​​​​
lvm LVM2 工​​​​​​​具​​​​​​​。​​​​​​​为​​​​​​​ LVM2 提​​​​​​​供​​​​​​​命​​​​​​​令​​​​​​​行​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​
system-config-lvm 为​​​​​​​ LVM2 提​​​​​​​供​​​​​​​图​​​​​​​形​​​​​​​化​​​​​​​用​​​​​​​户​​​​​​​界​​​​​​​面​​​​​​​。​​​​​​​
lvm.conf LVM 配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​。​​​​​​​完​​​​​​​整​​​​​​​路​​​​​​​径​​​​​​​是​​​​​​​ /etc/lvm/lvm.conf
群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​系​​​​​​​统​​​​​​​(Cluster Configuration System,CCS) ccs_tool ccs_tool 是​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​系​​​​​​​统​​​​​​​(CCS)的​​​​​​​一​​​​​​​部​​​​​​​分​​​​​​​。​​​​​​​它​​​​​​​被​​​​​​​用​​​​​​​来​​​​​​​制​​​​​​​作​​​​​​​ CSS 配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​的​​​​​​​在​​​​​​​线​​​​​​​更​​​​​​​新​​​​​​​。​​​​​​​此​​​​​​​外​​​​​​​,可​​​​​​​用​​​​​​​它​​​​​​​将​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​从​​​​​​​使​​​​​​​用​​​​​​​ GFS 6.0(或​​​​​​​更​​​​​​​早​​​​​​​版​​​​​​​本​​​​​​​)创​​​​​​​建​​​​​​​的​​​​​​​ CCS 归​​​​​​​档​​​​​​​升​​​​​​​级​​​​​​​为​​​​​​​ Red Hat Cluster Suite 发​​​​​​​行​​​​​​​版​​​​​​​本​​​​​​​中​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​ XML 格​​​​​​​式​​​​​​​。​​​​​​​
ccs_test 通​​​​​​​过​​​​​​​ ccsd 从​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​获​​​​​​​取​​​​​​​信​​​​​​​息​​​​​​​的​​​​​​​诊​​​​​​​断​​​​​​​和​​​​​​​测​​​​​​​试​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
ccsd 在​​​​​​​所​​​​​​​有​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​中​​​​​​​运​​​​​​​行​​​​​​​并​​​​​​​为​​​​​​​群​​​​​​​集​​​​​​​软​​​​​​​件​​​​​​​提​​​​​​​供​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​数​​​​​​​据​​​​​​​的​​​​​​​ CCS 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​。​​​​​​​
cluster.conf 这​​​​​​​个​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​。​​​​​​​完​​​​​​​整​​​​​​​的​​​​​​​路​​​​​​​径​​​​​​​是​​​​​​​ /etc/cluster/cluster.conf。​​​​​​​
群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​(Cluster Manager,CMAN) cman.ko CMAN 的​​​​​​​内​​​​​​​核​​​​​​​模​​​​​​​块​​​​​​​。​​​​​​​
cman_tool 这​​​​​​​是​​​​​​​ CMAN 的​​​​​​​管​​​​​​​理​​​​​​​前​​​​​​​台​​​​​​​。​​​​​​​它​​​​​​​启​​​​​​​动​​​​​​​和​​​​​​​停​​​​​​​止​​​​​​​ CMAN 并​​​​​​​能​​​​​​​修​​​​​​​改​​​​​​​某​​​​​​​些​​​​​​​内​​​​​​​部​​​​​​​参​​​​​​​数​​​​​​​,如​​​​​​​ vote。​​​​​​​
dlm_controld 由​​​​​​​ cman 初​​​​​​​始​​​​​​​脚​​​​​​​本​​​​​​​启​​​​​​​动​​​​​​​的​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​,它​​​​​​​管​​​​​​​理​​​​​​​内​​​​​​​核​​​​​​​中​​​​​​​的​​​​​​​ dlm;用​​​​​​​户​​​​​​​不​​​​​​​可​​​​​​​直​​​​​​​接​​​​​​​使​​​​​​​用​​​​​​​。​​​​​​​
gfs_controld cman 初​​​​​​​始​​​​​​​脚​​​​​​​本​​​​​​​启​​​​​​​动​​​​​​​的​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​,它​​​​​​​管​​​​​​​理​​​​​​​内​​​​​​​核​​​​​​​中​​​​​​​的​​​​​​​ gfs;用​​​​​​​户​​​​​​​不​​​​​​​可​​​​​​​直​​​​​​​接​​​​​​​使​​​​​​​用​​​​​​​。​​​​​​​
group_tool 用​​​​​​​来​​​​​​​获​​​​​​​取​​​​​​​与​​​​​​​保​​​​​​​护​​​​​​​(fencing)、​​​​​​​DLM、​​​​​​​GFS 相​​​​​​​关​​​​​​​的​​​​​​​组​​​​​​​群​​​​​​​列​​​​​​​表​​​​​​​,并​​​​​​​获​​​​​​​取​​​​​​​调​​​​​​​试​​​​​​​信​​​​​​​息​​​​​​​,包​​​​​​​括​​​​​​​ cman_tool services 在​​​​​​​ RHEL 4 中​​​​​​​提​​​​​​​供​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
groupd cman 初​​​​​​​始​​​​​​​脚​​​​​​​本​​​​​​​启​​​​​​​动​​​​​​​的​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​,它​​​​​​​是​​​​​​​ openais/cman 和​​​​​​​ dlm_controld/gfs_controld/fenced 之​​​​​​​间​​​​​​​的​​​​​​​接​​​​​​​口​​​​​​​;用​​​​​​​户​​​​​​​不​​​​​​​可​​​​​​​直​​​​​​​接​​​​​​​使​​​​​​​用​​​​​​​。​​​​​​​
libcman.so.<version number> 需​​​​​​​要​​​​​​​和​​​​​​​ cman.ko 交​​​​​​​互​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​所​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​库​​​​​​​。​​​​​​​
资​​​​​​​源​​​​​​​组​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​(Resource Group Manager,rgmanager) clusvcadm 手​​​​​​​动​​​​​​​启​​​​​​​用​​​​​​​、​​​​​​​禁​​​​​​​用​​​​​​​、​​​​​​​迁​​​​​​​移​​​​​​​和​​​​​​​重​​​​​​​启​​​​​​​群​​​​​​​集​​​​​​​中​​​​​​​用​​​​​​​户​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
clustat 用​​​​​​​来​​​​​​​显​​​​​​​示​​​​​​​群​​​​​​​集​​​​​​​状​​​​​​​态​​​​​​​(包​​​​​​​括​​​​​​​节​​​​​​​点​​​​​​​成​​​​​​​员​​​​​​​资​​​​​​​格​​​​​​​和​​​​​​​运​​​​​​​行​​​​​​​的​​​​​​​服​​​​​​​务​​​​​​​)的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
clurgmgrd 用​​​​​​​来​​​​​​​处​​​​​​​理​​​​​​​用​​​​​​​户​​​​​​​服​​​​​​​务​​​​​​​请​​​​​​​求​​​​​​​,如​​​​​​​服​​​​​​​务​​​​​​​启​​​​​​​动​​​​​​​、​​​​​​​服​​​​​​​务​​​​​​​禁​​​​​​​用​​​​​​​、​​​​​​​服​​​​​​​务​​​​​​​迁​​​​​​​移​​​​​​​和​​​​​​​服​​​​​​​务​​​​​​​重​​​​​​​启​​​​​​​的​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​。​​​​​​​
clurmtabd 处​​​​​​​理​​​​​​​群​​​​​​​集​​​​​​​的​​​​​​​ NFS 挂​​​​​​​载​​​​​​​表​​​​​​​的​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​。​​​​​​​
保​​​​​​​护​​​​​​​(Fence) fence_apc 用​​​​​​​于​​​​​​​ APC 电​​​​​​​源​​​​​​​开​​​​​​​关​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_bladecenter 用​​​​​​​于​​​​​​​带​​​​​​​有​​​​​​​ Telnet 接​​​​​​​口​​​​​​​的​​​​​​​ IBM Bladecenter 的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_bullpap 用​​​​​​​于​​​​​​​ Bull Novascale 平​​​​​​​台​​​​​​​管​​​​​​​理​​​​​​​处​​​​​​​理​​​​​​​器​​​​​​​(Platform Administration Processor,PAP)接​​​​​​​口​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_drac 用​​​​​​​于​​​​​​​ Dell 远​​​​​​​程​​​​​​​访​​​​​​​问​​​​​​​卡​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_ipmilan 用​​​​​​​于​​​​​​​通​​​​​​​过​​​​​​​ LAN 使​​​​​​​用​​​​​​​智​​​​​​​能​​​​​​​平​​​​​​​台​​​​​​​管​​​​​​​理​​​​​​​界​​​​​​​面​​​​​​​(Intelligent Platform Management Interface,IPMI)进​​​​​​​行​​​​​​​控​​​​​​​制​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_wti 用​​​​​​​于​​​​​​​ WTI 电​​​​​​​源​​​​​​​开​​​​​​​关​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_brocade 用​​​​​​​于​​​​​​​ Brocade 光​​​​​​​线​​​​​​​通​​​​​​​道​​​​​​​开​​​​​​​关​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_mcdata 用​​​​​​​于​​​​​​​ McData 光​​​​​​​线​​​​​​​通​​​​​​​道​​​​​​​开​​​​​​​关​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_vixel 用​​​​​​​于​​​​​​​ Vixel 光​​​​​​​线​​​​​​​通​​​​​​​道​​​​​​​开​​​​​​​关​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_sanbox2 用​​​​​​​于​​​​​​​ SANBox2 光​​​​​​​线​​​​​​​通​​​​​​​道​​​​​​​开​​​​​​​关​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_ilo 用​​​​​​​于​​​​​​​ HP ILO 接​​​​​​​口​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​(以​​​​​​​前​​​​​​​是​​​​​​​ fence_rib)。​​​​​​​
fence_rsa 用​​​​​​​于​​​​​​​ IBM RSA II 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_gnbd 用​​​​​​​于​​​​​​​ GNBD 存​​​​​​​储​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_scsi 用​​​​​​​于​​​​​​​ SCSI 持​​​​​​​久​​​​​​​保​​​​​​​留​​​​​​​的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_egenera 用​​​​​​​于​​​​​​​ Egenera BladeFrame 系​​​​​​​统​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_manual 手​​​​​​​动​​​​​​​交​​​​​​​互​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​注​​​​​​​意​​​​​​​:产​​​​​​​品​​​​​​​环​​​​​​​境​​​​​​​不​​​​​​​支​​​​​​​持​​​​​​​这​​​​​​​个​​​​​​​组​​​​​​​件​​​​​​​。​​​​​​​
fence_ack_manual fence_manual 代​​​​​​​理​​​​​​​的​​​​​​​用​​​​​​​户​​​​​​​界​​​​​​​面​​​​​​​。​​​​​​​
fence_node 在​​​​​​​单​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​上​​​​​​​执​​​​​​​行​​​​​​​ I/O 保​​​​​​​护​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​。​​​​​​​
fence_xvm 用​​​​​​​于​​​​​​​ Xen 虚​​​​​​​拟​​​​​​​机​​​​​​​的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​。​​​​​​​
fence_xvmd 用​​​​​​​于​​​​​​​ Xen 虚​​​​​​​拟​​​​​​​机​​​​​​​的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​宿​​​​​​​主​​​​​​​。​​​​​​​
fence_tool 加​​​​​​​入​​​​​​​和​​​​​​​离​​​​​​​开​​​​​​​保​​​​​​​护​​​​​​​域​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​。​​​​​​​
fenced I/O 保​​​​​​​护​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​。​​​​​​​
DLM libdlm.so.<version number> 支​​​​​​​持​​​​​​​分​​​​​​​布​​​​​​​式​​​​​​​锁​​​​​​​管​​​​​​​理​​​​​​​器​​​​​​​(Distributed Lock Manager,DLM)的​​​​​​​库​​​​​​​。​​​​​​​
GFS gfs.ko 在​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​使​​​​​​​用​​​​​​​的​​​​​​​内​​​​​​​核​​​​​​​模​​​​​​​块​​​​​​​,并​​​​​​​在​​​​​​​ GFS 群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​中​​​​​​​载​​​​​​​入​​​​​​​。​​​​​​​
gfs_fsck 修​​​​​​​复​​​​​​​未​​​​​​​挂​​​​​​​载​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
gfs_grow 扩​​​​​​​展​​​​​​​挂​​​​​​​载​​​​​​​的​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
gfs_jadd 在​​​​​​​已​​​​​​​挂​​​​​​​载​​​​​​​的​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​里​​​​​​​添​​​​​​​加​​​​​​​日​​​​​​​志​​​​​​​的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
gfs_mkfs 在​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​中​​​​​​​创​​​​​​​建​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
gfs_quota 在​​​​​​​已​​​​​​​挂​​​​​​​载​​​​​​​的​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​中​​​​​​​管​​​​​​​理​​​​​​​配​​​​​​​额​​​​​​​的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
gfs_tool 配​​​​​​​置​​​​​​​或​​​​​​​调​​​​​​​整​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​这​​​​​​​个​​​​​​​命​​​​​​​令​​​​​​​也​​​​​​​可​​​​​​​以​​​​​​​收​​​​​​​集​​​​​​​文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​各​​​​​​​种​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
mount.gfs mount(8) 所​​​​​​​调​​​​​​​用​​​​​​​的​​​​​​​挂​​​​​​​载​​​​​​​辅​​​​​​​助​​​​​​​工​​​​​​​具​​​​​​​;用​​​​​​​户​​​​​​​不​​​​​​​可​​​​​​​直​​​​​​​接​​​​​​​使​​​​​​​用​​​​​​​。​​​​​​​
GNBD gnbd.ko 在​​​​​​​客​​​​​​​户​​​​​​​端​​​​​​​中​​​​​​​使​​​​​​​用​​​​​​​ GNBD 设​​​​​​​备​​​​​​​驱​​​​​​​动​​​​​​​程​​​​​​​序​​​​​​​的​​​​​​​内​​​​​​​核​​​​​​​模​​​​​​​块​​​​​​​。​​​​​​​
gnbd_export 在​​​​​​​ GNBD 服​​​​​​​务​​​​​​​器​​​​​​​中​​​​​​​创​​​​​​​建​​​​​​​、​​​​​​​导​​​​​​​出​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​ GNBD 的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
gnbd_import 在​​​​​​​ GNBD 客​​​​​​​户​​​​​​​中​​​​​​​导​​​​​​​入​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​ GNBD 的​​​​​​​命​​​​​​​令​​​​​​​。​​​​​​​
gnbd_serv 允​​​​​​​许​​​​​​​节​​​​​​​点​​​​​​​通​​​​​​​过​​​​​​​网​​​​​​​络​​​​​​​导​​​​​​​出​​​​​​​本​​​​​​​地​​​​​​​存​​​​​​​储​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​。​​​​​​​
LVS pulse This is the controlling process which starts all other daemons related to LVS routers. At boot time, the daemon is started by the /etc/rc.d/init.d/pulse script. It then reads the configuration file /etc/sysconfig/ha/lvs.cf. On the active LVS router, pulse starts the LVS daemon. On the backup router, pulse determines the health of the active router by executing a simple heartbeat at a user-configurable interval. If the active LVS router fails to respond after a user-configurable interval, it initiates failover. During failover, pulse on the backup LVS router instructs the pulse daemon on the active LVS router to shut down all LVS services, starts the send_arp program to reassign the floating IP addresses to the backup LVS router's MAC address, and starts the lvs daemon.
lvsd 一​​​​​​​旦​​​​​​​被​​​​​​​ pulse 调​​​​​​​用​​​​​​​,lvs 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​就​​​​​​​会​​​​​​​在​​​​​​​活​​​​​​​跃​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​中​​​​​​​运​​​​​​​行​​​​​​​。​​​​​​​它​​​​​​​读​​​​​​​取​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​ /etc/sysconfig/ha/lvs.cf,调​​​​​​​用​​​​​​​ ipvsadm 工​​​​​​​具​​​​​​​来​​​​​​​构​​​​​​​建​​​​​​​和​​​​​​​维​​​​​​​护​​​​​​​ IPVS 路​​​​​​​由​​​​​​​表​​​​​​​,并​​​​​​​为​​​​​​​每​​​​​​​个​​​​​​​配​​​​​​​置​​​​​​​的​​​​​​​ LVS 服​​​​​​​务​​​​​​​分​​​​​​​配​​​​​​​ nanny 过​​​​​​​程​​​​​​​。​​​​​​​如​​​​​​​果​​​​​​​ nanny 报​​​​​​​告​​​​​​​某​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​关​​​​​​​闭​​​​​​​了​​​​​​​,lvs 将​​​​​​​指​​​​​​​引​​​​​​​ ipvsadm 工​​​​​​​具​​​​​​​从​​​​​​​ IPVS 路​​​​​​​由​​​​​​​表​​​​​​​中​​​​​​​删​​​​​​​除​​​​​​​这​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​。​​​​​​​
ipvsadm 这​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​更​​​​​​​新​​​​​​​内​​​​​​​核​​​​​​​中​​​​​​​的​​​​​​​ IPVS 路​​​​​​​由​​​​​​​表​​​​​​​。​​​​​​​lvs 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​通​​​​​​​过​​​​​​​调​​​​​​​用​​​​​​​ ipvsadm 添​​​​​​​加​​​​​​​、​​​​​​​修​​​​​​​改​​​​​​​或​​​​​​​删​​​​​​​除​​​​​​​ IPVS 路​​​​​​​由​​​​​​​表​​​​​​​里​​​​​​​的​​​​​​​条​​​​​​​目​​​​​​​来​​​​​​​设​​​​​​​立​​​​​​​并​​​​​​​管​​​​​​​理​​​​​​​ LVS。​​​​​​​
nanny nanny 监​​​​​​​控​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​在​​​​​​​活​​​​​​​跃​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​中​​​​​​​运​​​​​​​行​​​​​​​。​​​​​​​通​​​​​​​过​​​​​​​这​​​​​​​个​​​​​​​进​​​​​​​程​​​​​​​,活​​​​​​​跃​​​​​​​的​​​​​​​ LVS 路​​​​​​​由​​​​​​​器​​​​​​​可​​​​​​​以​​​​​​​决​​​​​​​定​​​​​​​每​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​的​​​​​​​健​​​​​​​康​​​​​​​状​​​​​​​况​​​​​​​,或​​​​​​​者​​​​​​​监​​​​​​​控​​​​​​​其​​​​​​​工​​​​​​​作​​​​​​​负​​​​​​​载​​​​​​​。​​​​​​​单​​​​​​​独​​​​​​​的​​​​​​​进​​​​​​​程​​​​​​​将​​​​​​​用​​​​​​​于​​​​​​​每​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​上​​​​​​​定​​​​​​​义​​​​​​​的​​​​​​​每​​​​​​​个​​​​​​​服​​​​​​​务​​​​​​​。​​​​​​​
lvs.cf 这​​​​​​​是​​​​​​​ LVS 的​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​。​​​​​​​其​​​​​​​完​​​​​​​整​​​​​​​路​​​​​​​径​​​​​​​是​​​​​​​ /etc/sysconfig/ha/lvs.cf。​​​​​​​所​​​​​​​有​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​都​​​​​​​直​​​​​​​接​​​​​​​或​​​​​​​间​​​​​​​接​​​​​​​地​​​​​​​从​​​​​​​这​​​​​​​个​​​​​​​文​​​​​​​件​​​​​​​获​​​​​​​取​​​​​​​配​​​​​​​置​​​​​​​信​​​​​​​息​​​​​​​。​​​​​​​
Piranha Configuration Tool 这​​​​​​​个​​​​​​​用​​​​​​​于​​​​​​​监​​​​​​​控​​​​​​​、​​​​​​​配​​​​​​​置​​​​​​​和​​​​​​​管​​​​​​​理​​​​​​​ LVS 的​​​​​​​基​​​​​​​于​​​​​​​ web 的​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​它​​​​​​​是​​​​​​​维​​​​​​​护​​​​​​​ /etc/sysconfig/ha/lvs.cf LVS 配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​的​​​​​​​默​​​​​​​认​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​
send_arp 在​​​​​​​故​​​​​​​障​​​​​​​切​​​​​​​换​​​​​​​过​​​​​​​程​​​​​​​中​​​​​​​,当​​​​​​​浮​​​​​​​动​​​​​​​ IP 地​​​​​​​址​​​​​​​从​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​更​​​​​​​改​​​​​​​为​​​​​​​另​​​​​​​一​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​时​​​​​​​,这​​​​​​​个​​​​​​​程​​​​​​​序​​​​​​​发​​​​​​​送​​​​​​​ ARP 广​​​​​​​播​​​​​​​。​​​​​​​
Quorum 磁​​​​​​​盘​​​​​​​ qdisk 用​​​​​​​于​​​​​​​ CMAN / Linux-Cluster 的​​​​​​​基​​​​​​​于​​​​​​​磁​​​​​​​盘​​​​​​​的​​​​​​​ quorum 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​。​​​​​​​
mkqdisk 群​​​​​​​集​​​​​​​ Quorum 磁​​​​​​​盘​​​​​​​工​​​​​​​具​​​​​​​。​​​​​​​
qdiskd 群​​​​​​​集​​​​​​​ Quorum 磁​​​​​​​盘​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​。​​​​​​​

2.2. Man Page

下​​​​​​​面​​​​​​​的​​​​​​​内​​​​​​​容​​​​​​​列​​​​​​​出​​​​​​​了​​​​​​​和​​​​​​​ Red Hat Cluster Suite 相​​​​​​​关​​​​​​​的​​​​​​​ man page 以​​​​​​​及​​​​​​​其​​​​​​​它​​​​​​​资​​​​​​​源​​​​​​​。​​​​​​​
  • 群​​​​​​​集​​​​​​​基​​​​​​​础​​​​​​​结​​​​​​​构​​​​​​​
    • ccs_tool (8) - 制​​​​​​​作​​​​​​​ CCS 配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​的​​​​​​​在​​​​​​​线​​​​​​​更​​​​​​​新​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​
    • ccs_test (8) - 诊​​​​​​​断​​​​​​​运​​​​​​​行​​​​​​​的​​​​​​​群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​系​​​​​​​统​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​
    • ccsd (8) - 访​​​​​​​问​​​​​​​ CCS 群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​的​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
    • ccs (7) - 群​​​​​​​集​​​​​​​配​​​​​​​置​​​​​​​系​​​​​​​统​​​​​​​
    • cman_tool (8) - 群​​​​​​​集​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​
    • cluster.conf [cluster] (5) - 群​​​​​​​集​​​​​​​产​​​​​​​品​​​​​​​的​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​
    • qdisk (5) - 用​​​​​​​于​​​​​​​ CMAN / Linux-Cluster 的​​​​​​​基​​​​​​​于​​​​​​​磁​​​​​​​盘​​​​​​​的​​​​​​​ quorum 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
    • mkqdisk (8) - 群​​​​​​​集​​​​​​​ Quorum 磁​​​​​​​盘​​​​​​​工​​​​​​​具​​​​​​​
    • qdiskd (8) - 群​​​​​​​集​​​​​​​ Quorum 磁​​​​​​​盘​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
    • fence_ack_manual (8) - 作​​​​​​​为​​​​​​​手​​​​​​​工​​​​​​​ I/O 保​​​​​​​护​​​​​​​的​​​​​​​一​​​​​​​部​​​​​​​分​​​​​​​运​​​​​​​行​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​
    • fence_apc (8) - APC MasterSwitch 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_bladecenter (8) - IBM Bladecenter 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_brocade (8) - IBrocade FC switch 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_bullpap (8) - PAP 管​​​​​​​理​​​​​​​控​​​​​​​制​​​​​​​台​​​​​​​控​​​​​​​制​​​​​​​的​​​​​​​ Bull FAME 架​​​​​​​构​​​​​​​的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_drac (8) - Dell 远​​​​​​​程​​​​​​​访​​​​​​​问​​​​​​​卡​​​​​​​的​​​​​​​保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_egenera (8) - Egenera BladeFrame 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_gnbd (8) - 基​​​​​​​于​​​​​​​ GNBD 的​​​​​​​ GFS 群​​​​​​​集​​​​​​​的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_ilo (8) - HP Integrated Lights Out card 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_ipmilan (8) - IPMI 通​​​​​​​过​​​​​​​局​​​​​​​域​​​​​​​网​​​​​​​控​​​​​​​制​​​​​​​的​​​​​​​机​​​​​​​器​​​​​​​的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_manual (8) - 作​​​​​​​为​​​​​​​手​​​​​​​工​​​​​​​ I/O 保​​​​​​​护​​​​​​​的​​​​​​​一​​​​​​​部​​​​​​​分​​​​​​​运​​​​​​​行​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​
    • fence_mcdata (8) - 用​​​​​​​于​​​​​​​ McData FC switch 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_node (8) - 在​​​​​​​单​​​​​​​个​​​​​​​节​​​​​​​点​​​​​​​上​​​​​​​执​​​​​​​行​​​​​​​ I/O 保​​​​​​​护​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​
    • fence_rib (8) - 用​​​​​​​于​​​​​​​ Compaq Remote Insight Lights Out card 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_rsa (8) - IBM RSA II 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_sanbox2 (8) - 用​​​​​​​于​​​​​​​ QLogic SANBox2 FC switch 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_scsi (8) - 用​​​​​​​于​​​​​​​ SCSI persistent reservation 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_tool (8) - 加​​​​​​​入​​​​​​​和​​​​​​​离​​​​​​​开​​​​​​​保​​​​​​​护​​​​​​​域​​​​​​​的​​​​​​​程​​​​​​​序​​​​​​​
    • fence_vixel (8) - 用​​​​​​​于​​​​​​​ Vixel FC switch 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_wti (8) - 用​​​​​​​于​​​​​​​ WTI Network Power Switch 的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_xvm (8) - 用​​​​​​​于​​​​​​​ Xen 虚​​​​​​​拟​​​​​​​机​​​​​​​的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​
    • fence_xvmd (8) - 用​​​​​​​于​​​​​​​ Xen 虚​​​​​​​拟​​​​​​​机​​​​​​​的​​​​​​​ I/O 保​​​​​​​护​​​​​​​代​​​​​​​理​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
    • fenced (8) - I/O 保​​​​​​​护​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
  • 高​​​​​​​可​​​​​​​用​​​​​​​性​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​
    • clusvcadm (8) - 群​​​​​​​集​​​​​​​用​​​​​​​户​​​​​​​服​​​​​​​务​​​​​​​管​​​​​​​理​​​​​​​工​​​​​​​具​​​​​​​
    • clustat (8) - 检​​​​​​​查​​​​​​​群​​​​​​​集​​​​​​​状​​​​​​​态​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​
    • Clurgmgrd [clurgmgrd] (8) - 资​​​​​​​源​​​​​​​组​​​​​​​(群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​)管​​​​​​​理​​​​​​​者​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
    • clurmtabd (8) - 群​​​​​​​集​​​​​​​ NFS 远​​​​​​​程​​​​​​​挂​​​​​​​载​​​​​​​表​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
  • GFS
    • gfs_fsck (8) - 脱​​​​​​​机​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​检​​​​​​​查​​​​​​​程​​​​​​​序​​​​​​​
    • gfs_grow (8) - 扩​​​​​​​展​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​
    • gfs_jadd (8) - 在​​​​​​​ GFS 文​​​​​​​件​​​​​​​系​​​​​​​统​​​​​​​里​​​​​​​添​​​​​​​加​​​​​​​日​​​​​​​志​​​​​​​(jounal)
    • gfs_mount (8) - GFS 挂​​​​​​​载​​​​​​​选​​​​​​​项​​​​​​​
    • gfs_quota (8) - 控​​​​​​​制​​​​​​​ GFS 磁​​​​​​​盘​​​​​​​配​​​​​​​额​​​​​​​(quota)
    • gfs_tool (8) - GFS ioctl 调​​​​​​​用​​​​​​​接​​​​​​​口​​​​​​​
  • 群​​​​​​​集​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​者​​​​​​​(Cluster Logical Volume Manager)
    • clvmd (8) - 群​​​​​​​集​​​​​​​ LVM 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
    • lvm (8) - LVM2 工​​​​​​​具​​​​​​​
    • lvm.conf [lvm] (5) - LVM2 的​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​
    • lvmchange (8) - 修​​​​​​​改​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​管​​​​​​​理​​​​​​​者​​​​​​​的​​​​​​​属​​​​​​​性​​​​​​​
    • pvcreate (8) - 初​​​​​​​始​​​​​​​化​​​​​​​ LVM 使​​​​​​​用​​​​​​​的​​​​​​​磁​​​​​​​盘​​​​​​​或​​​​​​​分​​​​​​​区​​​​​​​
    • lvs (8) - 报​​​​​​​告​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​
  • 全​​​​​​​球​​​​​​​网​​​​​​​络​​​​​​​块​​​​​​​设​​​​​​​备​​​​​​​(Global Network Block Device)
    • gnbd_export (8) - 导​​​​​​​出​​​​​​​ GNBD 的​​​​​​​接​​​​​​​口​​​​​​​
    • gnbd_import (8) - 操​​​​​​​纵​​​​​​​客​​​​​​​户​​​​​​​机​​​​​​​上​​​​​​​的​​​​​​​ GNBD 块​​​​​​​设​​​​​​​备​​​​​​​
    • gnbd_serv (8) - gnbd 服​​​​​​​务​​​​​​​器​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
  • LVS
    • pulse (8) - 监​​​​​​​控​​​​​​​群​​​​​​​集​​​​​​​节​​​​​​​点​​​​​​​健​​​​​​​康​​​​​​​状​​​​​​​况​​​​​​​的​​​​​​​ heartbeating 守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
    • lvs.cf [lvs] (5) - lvs 的​​​​​​​配​​​​​​​置​​​​​​​文​​​​​​​件​​​​​​​
    • lvscan (8) - 扫​​​​​​​描​​​​​​​(所​​​​​​​有​​​​​​​磁​​​​​​​盘​​​​​​​)的​​​​​​​逻​​​​​​​辑​​​​​​​卷​​​​​​​。​​​​​​​
    • lvsd (8) - 控​​​​​​​制​​​​​​​红​​​​​​​帽​​​​​​​群​​​​​​​集​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​守​​​​​​​护​​​​​​​进​​​​​​​程​​​​​​​
    • ipvsadm (8) - Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​管​​​​​​​理​​​​​​​
    • ipvsadm-restore (8) - 从​​​​​​​标​​​​​​​准​​​​​​​输​​​​​​​入​​​​​​​恢​​​​​​​复​​​​​​​ IPVS 表​​​​​​​
    • ipvsadm-save (8) - 把​​​​​​​ IPVS 保​​​​​​​存​​​​​​​至​​​​​​​标​​​​​​​准​​​​​​​输​​​​​​​出​​​​​​​
    • nanny (8) - 监​​​​​​​控​​​​​​​群​​​​​​​集​​​​​​​里​​​​​​​服​​​​​​​务​​​​​​​的​​​​​​​状​​​​​​​态​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​
    • send_arp (8) - 将​​​​​​​新​​​​​​​的​​​​​​​ IP 地​​​​​​​址​​​​​​​/MAC 地​​​​​​​址​​​​​​​映​​​​​​​射​​​​​​​通​​​​​​​知​​​​​​​网​​​​​​​络​​​​​​​的​​​​​​​工​​​​​​​具​​​​​​​

2.3. 兼​​​​​​​容​​​​​​​的​​​​​​​硬​​​​​​​件​​​​​​​

有​​​​​​​关​​​​​​​与​​​​​​​Red Hat Cluster Suite组​​​​​​​件​​​​​​​兼​​​​​​​容​​​​​​​的​​​​​​​硬​​​​​​​件​​​​​​​的​​​​​​​信​​​​​​​息​​​​​​​(例​​​​​​​如​​​​​​​:支​​​​​​​持​​​​​​​保​​​​​​​护​​​​​​​设​​​​​​​备​​​​​​​、​​​​​​​存​​​​​​​储​​​​​​​设​​​​​​​备​​​​​​​和​​​​​​​光​​​​​​​纤​​​​​​​交​​​​​​​换​​​​​​​机​​​​​​​),请​​​​​​​参​​​​​​​考​​​​​​​硬​​​​​​​件​​​​​​​配​​​​​​​置​​​​​​​指​​​​​​​南​​​​​​​,地​​​​​​​址​​​​​​​为​​​​​​​ http://www.redhat.com/cluster_suite/hardware/。​​​​​​​

附录 A. 修​​​​​​​订​​​​​​​记​​​​​​​录​​​​​​​

修订历史
修订 3-8.4002013-10-31Rüdiger Landmann
Rebuild with publican 4.0.0
修订 3-82012-07-18Anthony Towns
Rebuild for Publican 3.0
修订 1.0-0Tue Jan 20 2008Paul Kennedy
统​​​​​​​一​​​​​​​发​​​​​​​布​​​​​​​重​​​​​​​点​​​​​​​

索引

M

man pages
cluster components,Man Page

N

NAT
routing methods, LVS,路​​​​​​​由​​​​​​​方​​​​​​​法​​​​​​​
network address translation (见 NAT)

O

overview
economy,Red Hat GFS
performance,Red Hat GFS
scalability,Red Hat GFS

P

Piranha Configuration Tool
CONTROL/MONITORING,CONTROL/MONITORING
EDIT MONITORING SCRIPTS Subsection,EDIT MONITORING SCRIPTS Subsection
GLOBAL SETTINGS,GLOBAL SETTINGS
login panel,Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​管​​​​​​​理​​​​​​​图​​​​​​​形​​​​​​​用​​​​​​​户​​​​​​​界​​​​​​​面​​​​​​​
necessary software,Linux 虚​​​​​​​拟​​​​​​​服​​​​​​​务​​​​​​​器​​​​​​​管​​​​​​​理​​​​​​​图​​​​​​​形​​​​​​​用​​​​​​​户​​​​​​​界​​​​​​​面​​​​​​​
REAL SERVER subsection,「​​​​​​​REAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​
REDUNDANCY,REDUNDANCY
VIRTUAL SERVER subsection,VIRTUAL SERVERS
Firewall Mark ,「​​​​​​​VIRTUAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​
Persistence ,「​​​​​​​VIRTUAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​
Scheduling ,「​​​​​​​VIRTUAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​
Virtual IP Address ,「​​​​​​​VIRTUAL SERVER」​​​​​​​分​​​​​​​支​​​​​​​
VIRTUAL SERVERS,VIRTUAL SERVERS

法律通告

Copyright © 2009 Red Hat, Inc.
This document is licensed by Red Hat under the Creative Commons Attribution-ShareAlike 3.0 Unported License. If you distribute this document, or a modified version of it, you must provide attribution to Red Hat, Inc. and provide a link to the original. If the document is modified, all Red Hat trademarks must be removed.
Red Hat, as the licensor of this document, waives the right to enforce, and agrees not to assert, Section 4d of CC-BY-SA to the fullest extent permitted by applicable law.
Red Hat, Red Hat Enterprise Linux, the Shadowman logo, JBoss, OpenShift, Fedora, the Infinity logo, and RHCE are trademarks of Red Hat, Inc., registered in the United States and other countries.
Linux® is the registered trademark of Linus Torvalds in the United States and other countries.
Java® is a registered trademark of Oracle and/or its affiliates.
XFS® is a trademark of Silicon Graphics International Corp. or its subsidiaries in the United States and/or other countries.
MySQL® is a registered trademark of MySQL AB in the United States, the European Union and other countries.
Node.js® is an official trademark of Joyent. Red Hat Software Collections is not formally related to or endorsed by the official Joyent Node.js open source or commercial project.
The OpenStack® Word Mark and OpenStack logo are either registered trademarks/service marks or trademarks/service marks of the OpenStack Foundation, in the United States and other countries and are used with the OpenStack Foundation's permission. We are not affiliated with, endorsed or sponsored by the OpenStack Foundation, or the OpenStack community.
All other trademarks are the property of their respective owners.

为了尽快向用户提供最新的信息,本文档可能会包括由机器自动从英文原文翻译的内容。如需更多信息,请参阅此说明。