How to deploy IPMP

nimysun

In order to setup a failover configuration for IPMP I would recommend doing it via this way around:
create some relative configuration files such as /etc/hostname.NICn and so on;
/etc/hostname.ce0
RIP01 deprecated -failover netmask + broadcast + group group1 up \
addif VIP01 netmask + broadcast + up
/etc/hostname.ce1
RIP02 deprecated -failover netmask + broadcast + standby group group1 up
/etc/hosts
10.10.10.1 RIP01
10.10.10.2 RIP02
10.10.10.3 VIP01 #this is the ip the users will connect to
/etc/defaultrouter
router's ip
Now reboot servers
Notes :
a) you don't want the ip's you assign to the physical interfaces to failover. configure them as nofailover (-failover)
b) you don't want load spreading on both physical interfaces, so the VIP will only "belong" to the active interface. don't do another addif for the 2nd physical interface
c) you want the 2nd physical interface to take over if the first one fails. configure it as standby and nofailover (-failover)
To test this:
#ifconfig -a (should give u something like this)
ce0: flags=9040843 mtu 1500 index 2
inet 10.10.10.1 netmask fffffc00 broadcast 10.10.10.255
groupname grp1
ce0:1: flags=1000843 mtu 1500 index 2
inet 10.10.10.3 netmask fffffc00 broadcast 10.10.10.255
ce1: flags=9040843 mtu 1500 index 3
inet 10.10.10.2 netmask fffffc00 broadcast 10.10.10.255
groupname grp1
1. open a terminal window on another server
2. #ping -s 10.10.10.3
3. pull out the cable from ce0
4. the ping should continue with messages being echoed to the /var/adm/messages file along these lines :
in.mpathd[573]: [ID 215189 daemon.error] The link has gone down on ce0
in.mpathd[573]: [ID 594170 daemon.error] NIC failure detected on ce0 of group1
in.mpathd[573]: [ID 832587 daemon.error] Successfully failed over from NIC ce0 to NIC ce1
5. plug the cable back and there should be an auto-failover back to ce0


本文来自ChinaUnix博客，如果查看原文请点：http://blog.chinaunix.net/u/43/showart_96824.html