fallback takes all connections even though still 1 realserver available

[Thoralf Will <thoralf@xxxxxxxxxxx>]

Meanwhile I got my fallback problem solved and the webserver on 127.0.0.1 is 
answering. Interesting is that lvs ignores the give ip-address to connect to 
the webserver but uses the real ip. So the problem was the Listen that was 
restricted to 127.0.0.1:80 which couldn't work.

Ok, here is the problem:

root@server:~ > ipvsadm
IP Virtual Server version 1.0.11 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  192.168.1.110:http wlc persistent 60
  -> localhost:http               Local   1      0          12
  -> celeborn:http                Route   1      0          0

Why is the fallback (already) active (and answering all requests) even though 
there is still a working realserver online, capable to do answer the requests 
as expected?

configuration:

checktimeout=10
checkinterval=10
quiescent=yes

virtual = 192.168.1.110:80
        protocol = tcp
        service = http
        scheduler = wlc
        real = 192.168.1.200:80 gate
        real = 192.168.1.201:80 gate
        fallback = 127.0.0.1:80
        request = "test.html"
        receive = "teststring"
        persistent=60

Also interesting: the fallback remains active in the configuration, even when 
the webserver on the localhost is shut down (and therefor the whole system 
appears to be down, even though 1 realserver is still alive), the realservers 
get added and removed dynamically depending on their status.
When I remove the entry for the fallback everything is working fine - but I 
really want to have a fallback in case ALL realservers die.

What's wrong?

thx,
Thoralf