On Tue, Feb 12, 2013 at 1:32 PM, Vik Killa <vipkilla@gmail.com> wrote:
>
> I would like to see that statement put to the test. I would like to see
> someone attempt to put Corosync (or whatever existing solution) to use as
> the HA engine in a FS cluster. I suspect the amount of work involved is way
> more than has been suggested. If it really were that easy ("easy" being
> relative compared to what people perceive Eliot is doing) then I suspect it
> would have been done already.

Hey MC,
http://wiki.freeswitch.org/wiki/Enterprise_deployment_with_Corosync
^ I wrote that months ago. Pretty easy :)

ip link set <your_interface> arp off

Watch your cluster sit there like an idiot while your traffic fails and you spend hours trying to find the cause.

ethtool -s <your_interface> speed 10 duplex half

Watch your cluster sit there like an idiot while your traffic fails and you spend hours trying to find the cause.

Rinse and repeat with literally dozens upon dozens of other things you could do where your cluster just sits there broken and you pull your hair out.

Still seem simple?