|
On Sat, 2005-04-23 at 23:48 +0200, eternityos@xxxxxxx wrote:
> I decided to switch to keepalived, to have a failover
> so now i got a good healthcheck of nodes and when checking
> for the table, it is updated depending on if a service
> is available or not...
>
> ipvsadm -l shows this :
>
> node11:~# ipvsadm -l
> IP Virtual Server version 1.2.0 (size=4096)
> Prot LocalAddress:Port Scheduler Flags
> -> RemoteAddress:Port Forward Weight ActiveConn InActConn
> TCP 10.3.7.50:www rr persistent 10
> -> 192.168.1.13:www Masq 1 0 228
> -> 192.168.1.12:www Masq 1 0 0
> -> 192.168.1.11:www Local 1 0 0
> node11:~#
>
> As you can see, round robin is set, there is real servers and
> the real servers are well configured since if i stop the service
> on node 192.168.1.13, the next node start answering to the client.
> Now i am kind of stuck, LVS works on this machine because i got
> a load balancing with ipvsadm (with rr too). ipvsadm -l shows
> it is well configured apparently, so now, How can i investigate
> this ? Anyone knows about that stuff ?
I'm not entirely sure I understand what your problem is; however I see
you're using persistence and from the output it seems likely that all
connections are going to one server and not the others.
Possible solutions:
1. Stop using persistence.
2. Use a different persistence mask (or none at all).
3. Subscribe to the keepalived list :) and post your config there after
explaining what the problem is.
Graeme
|
