|
Chris, I am rather confused about why you are using
both heartbeat and ldirectord+LVS for this particular setup.
It sounds like you use one of the following solutions:
1. Heartbat (No LVS or ldirectord)
|-------------------------------|
| |
----------------- -----------------
|real server | |fallback server|
|10.0.0.1 | |10.0.0.2 |
----------------- -----------------
Heartbeat monitors the health of the real and fallback server.
It may start apache, or it may already be running on both machines.
Heartbeat manages the VIP 192.168.1.225 and will move
it to the failback-server (actually this should be called the sandby
in this case) if the real-server fails.
2. LVS and Ldirectord (no heartbeat)
----------------
|virtual server|
|192.168.1.225 |
----------------
|
|
|
|-------------------------------|
| |
----------------- -----------------
|real server | |fallback server|
|10.0.0.1 | |10.0.0.2 |
----------------- -----------------
Apache should be running on both the real server and the
fallback-server.
Ldirectord monitors the health of the real-server.
When it is ok, configures LVS to forward all traffic
to the real-server. When the real-server fails,
it configures LVS to forward all traffic to the fallback-server.
I note that you have chosen to use the "gate" (direct-routing)
method to forward your packets. This is somewhat tricky
to set up. If you are having troubles I would recommend
investingating references to the ARP problem in the LVS-HOWTO.
Actually, I would recommend using "masq" (NAT) if you
are starting out as it is a lot easier to get up and
running in my opinion.
--
Horms
|
