|
On Wed, Aug 13, 2003 at 12:09:16PM +0000, patrick kuah wrote:
> Hi,
>
> I had followed the doc from ultra monkey which horms sent to me [Thanks
> Horms :) ], it about how to install LVS on two servers and load balacing
> each other.
> http://www.ultramonkey.org/2.0.1/topologies/sl-ha-lb-eg.html
>
> I got two servers (web1 and web2)
> I had created eth0:0 and lo:0 on both of my servers, but when ever the
> linux restart,
> both eth0:0 and lo:0 come online at the same time which i belive it should
> not be...
No, that should not be.
> If i not wrong,
> The active server (web1) will have eth0:0 and the passive server (web2)
> will have lo:0 online, i need to restart the heartbeat service in order to
> achieve this.
> In my haresource file i only specify the active node (web1), VIP (eth0:0
> and lo:0) and the ldirectord as the resource on both servers.
> How to make sure whenever
No, you do not need to restart heartbeat. But make sure that
you have a recent version, I would recommend 1.0.3.
When you boot each system up make sure that you configre
lo:0 to have the VIP on both systems. Make sure this is
done before heartbeat is started. When heartbeat starts it
will remove the lo:0 inteface if it brings up the vip on eth0:0
and vice versa. But this only works if the VIP was on lo:0
in the first place.
If this does not work I would suggest running IPaddr manually from
the command line, perhaps under bash -x to trace what is going on.
This can be a bit tricky to get working, but it does work.
Please let me know if you find any bugs, there is always
room for improvement.
> Is my configuration correct ???
>
> In the ldirectord.cf file, what does the weight mean ???
See the ipvsadm man page for a description of how weights work.
--
Horms
|
