[Linux-cluster] Network detect failure

[KC LO]

Hi all,

I just configured a two node(node01, node02) cluster and the IP resources
defined with

<ip address="10.1.1.1" monitor_link="1"/>

I observed that the server will conduct ping test every 60 seconds.

Every day, the active node will get network link detect failure several
times.(The time and frequency happen randomly)
Nov 21 14:39:29 node02 clurgmgrd[5841]: <notice> status on ip "10.1.1.1"
returned 1 (generic error)

When node02 detected failure, it can successfully fail-over to node01.
However, the network link in node01 will also detect failure within several
hours.  It will then auto fail-over to node02.  It happens between node01
and node02.  I don't see any link down in /var/log/messages or network
switches.

Do you have any ideas?

Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20111121/1035328f/attachment.htm>