fallback doesn't work at all (apache/ha/lvs)

Thoralf Will thoralf at cipsoft.com
Fri Apr 23 11:41:53 BST 2004

I'm new to the list, sorry if this hasn't been answered before, I didn't find anything so far.

I have a (test)setup with 1 node (2nd node is just a fake-entry) as a director for lvs. The director is connected to 2 realservers which are working fine and doing their job. The problem is the fallback that I want to implement in case both servers die. This fallback (apache) is meant to run locally on the director. On the director it works fine, I can request the testpage and get the expected answer. But if I switch of both realservers and the fallback gets added to the list I don't receive the expected waitpage but a connection refused error. The director notices an incomming connection regarding to ipvsadm but no answer is delivered.

Any idea where the mistake maybe?

thx in advance,


/etc/ha.d/ha.cf: (shouldn't be the problem)
debugfile /var/log/ha-debug
logfile /var/log/ha-log
logfacility     local0
keepalive 2
deadtime 10
serial  /dev/ttyS0
baud    19200
udpport 694
bcast     eth0
auto_failback on
node    director.somedomain.ext
node    fakeserver.somedomain.ext

director.somedomain.ext ldirectord

/etc/ha.d/conf/ldirectord.cf: (maybe here somewhere?)
virtual =
        protocol = tcp
        scheduler = wlc
        real = gate 5
        real = gate 5
        fallback =
        request = "test.html"
        receive = "teststring"

output of ipvsadm:
IP Virtual Server version 1.0.11 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP wlc persistent 60
  -> realserver1.somedomain.ext:http   Route   0      0          0
  -> realserver2.somedomain.ext:http   Route   0      0          0
  -> localhost:http                    Local   1      0          0

(the fallback-webserver is listening on, the realservers are shut off to test the fallback)

More information about the lvs-users mailing list