From: Brian J. Watson <Brian.J.Watson@hp...> - 2004-12-22 09:46:12
Roger Tsang wrote:
>> I guess you can wait until 1.2.1 to include them. What they do is
>> allow a node to automatically reboot 15 seconds after it panics,
>> rather than drop into kdb and become unavailable from the cluster.
> For Fedora , you also need to pass kdb=off
My config file changes build this into the kernel.