Keepalived keeps forcing new elections despite priority not 255
chris at meetup.com
Tue Aug 30 20:01:03 BST 2005
Just trying to pick up on July's thread, as I hit the same problems with
Gentoo. Either server works fine, and failover works, in the same
sense, but when both servers are running, the backup (priority 100)
tries and fails to take over every second, flooding the logs and sending
out an announcement every second. The master has a priority of 150.
I noticed that Peter Broadwell had this problem, and got around it with
some clever iptables rules, but he never specified what those rules were.
Does anyone here have a guideline? tcpdump seems to see the packets
from either server going across, but I have to be missing something to
see this kind of behavior.
More information about the lvs-users