基于heartbeat的NFS HA实现

基于heartbeat的NFS HA实现

Version: 1.0
Maintainer: hutuworm
Update: 2003-09-30

※系统说明

    本系统由两台Dell PowerEdge 2650（HUTUNFS1 & HUTUNFS2）和一台磁盘阵列组成，磁盘阵列
划分为ld0 & ld1两个RAID5逻辑盘（Logical Disk）。操作系统为RedHat 7.3 Linux。设置NFS HA
的目的是，当HUTUNFS1失败时HUTUNFS2自动接管。

1 安装heartbeat 1.0.3 for RedHat 7.3
  http://www.ultramonkey.org/download/heartbeat/1.0.3/redhat_7.3/
  heartbeat-1.0.3-1.rh.7.3.1.i386.rpm
  heartbeat-ldirectord-1.0.3-1.rh.7.3.1.i386.rpm
  heartbeat-pils-1.0.3-1.rh.7.3.1.i386.rpm
  heartbeat-stonith-1.0.3-1.rh.7.3.1.i386.rpm 

2 设置配置文件:
  HUTUNFS1为master，HUTUNFS2为slave。
 
2.1 HUTUNFS1  192.168.0.30
    /etc/ha.d/ha.cf
      # define nodes in cluster
      node hutunfs1
      node hutunfs2
      # time a system must be unreachable before considered dead (seconds)
      deadtime 5
      # set up for the serial heartbeat pulse
      # 如果两台服务器之间用心跳线连接，heartbeat支持串行心跳监测。
      #serial /dev/ttyS0
      #baud 19200
      # interface to run the network heartbeat pulse
      udp bond0

    /etc/ha.d/authkeys
      auth 3
      3 md5 hutuworm

    /etc/ha.d/haresources
      hutunfs1 192.168.0.30 Filesystem::/dev/sdc1::/ld0::ext3 nfslock nfs
      hutunfs1 192.168.0.30 Filesystem::/dev/sdd1::/ld1::ext3 nfslock nfs

    /etc/exports
      #
      # Export the shared disk, allowing read/write access and
      # synchronous I/O with no write delay.
      /ld0 192.168.0.*(rw,sync,no_wdelay)
      /ld1 192.168.0.*(rw,sync,no_wdelay)

    /etc/hosts
      # Do not remove the following line, or various programs
      # that require network functionality will fail.
      127.0.0.1               hutunfs1 localhost.localdomain localhost
      192.168.0.30  hutunfs1
      192.168.0.31  hutunfs2

2.2 HUTUNFS2  192.168.0.31
    /etc/ha.d/ha.cf
    /etc/ha.d/authkeys
    /etc/ha.d/haresources
    /etc/exports
    以上四个文件与HUTUNFS1相同

    /etc/hosts
      # Do not remove the following line, or various programs
      # that require network functionality will fail.
      127.0.0.1               hutunfs2 localhost.localdomain localhost
      192.168.0.30           hutunfs1
      192.168.0.31           hutunfs2

3. 在HUTUNFS1 & HUTUNFS2上运行ntsysv，确认nfs服务关闭，heartbeat服务打开。
   先重起HUTUNFS1，再重起HUTUNFS2，HUTUNFS1作为master运行，HUTUNFS2作为slave运行。
   当HUTUNFS1失效时，HUTUNFS2将自动接管192.168.0.30，并同时开启NFS服务。