Hi All,
I have two boxes running:
Apache/Tomcat/JK2
heartbeat
ldirectord
(ie. I have collapsed the real servers into the directors).
heartbeat is working fine and fails over/back arp + ldirectord
ldirectord runs fine with the health checks and takes the real servers
on/offline etc.
Load balancing is round robin with equal weights.
The issue is that I can only access Apache on the real server where ldirectord
is active (I know the topology should have four machines but I don't have that
much hardware. Additionally I have tested a single director / single real
server setup using NAT fine).
I am not sure of the ldirectord setup that I should use in order to be able to
forward the HTTP request to the second machine (where ldirectord is passive).
Should I be using the NAT setup or one of the direct server return topologies
in this configuration.
Any help gratefully appreciated.
Here are my configs (sorry about the wrap):
haresources-
lvs01.sublime.com.au 192.168.1.2
ldirectord::/usr/local/etc/ha.d/conf/ldirectord.conf
ha.cf-
serial /dev/ttyS0
watchdog /dev/watchdog
bcast eth1
keepalive 2
warntime 10
initdead 120
#hopfudge 1
baud 19200
udpport 694
nice_failback on
node lvs01.sublime.com.au
node lvs02.sublime.com.au
#respawn hacluster /usr/local/lib/heartbeat/ipfail
#respawn hacluster /usr/local/lib/heartbeat/ccm
ping 192.168.1.1
logfile /var/log/ha-log
debugfile /var/log/ha-debug
ldirectord.conf
checktimeout = 10
checkinterval = 10
autoreload = yes
quiescent = yes
logfile = "/var/log/ldirectord.log"
virtual = 192.168.1.2:80
real = 127.0.0.1:80 gate 5
real = 192.168.1.12:80 gate 5
service = http
request =
"/examples/servlet/HelloWorldExample"
receive = "Hello World!"
scheduler = rr
protocol = tcp
Regards,
James
|