[Linux-HA] Normal failover time

[Guochun Shi]

At 03:10 PM 6/28/2004 +0100, you wrote:

>Hello,
>
>I've set up a simple cluster in Linux (RH) which only implements a new IP address. It takes about 90 secons to failover manually. Shouldn't it be faster? What's the problem?
>
>Here are the config files:
>
>ha.cf - node 1 and node 2
>=========================
>logfacility     local0
>keepalive 1
>deadtime 3
>warntime 3
>initdead 20
>udpport 694
>baud    19200
>serial  /dev/ttyS0      # Linux
>bcast   eth0            # Linux
>auto_failback off
>node prxnms1n
>node prxnms2n
>ping 10.100.22.1

Config ping without ipfail does not make much sense.
You may want to remove this line or add ipfail

-Guochun